[opensuse] Tape Drive Trouble
Hi, I have 4 tape drives in my server and I am lately experiencing a few problems. One night a backup failed with rmtio failed: Input output Error. I ran it again and it worked. This has happened more and more frequently and now the drive fails writing every time with this error. I put the drive (Hp storageworks Ultrium 920) in another server. It works fine. I have just received a storageworks ultrium 1760 drive today. Same problem is occuring. The easiest way to make the error come up is to try mt -f /dev/st3 erase The drive sounds like it erases the tape, then the tape is unloaded and goes back in again automatically. Then it erases it again, then does it again then again and again then eventually it gives up with the message mt: /dev/st3: rmtioctl failed: Input/output error Can anyone enlighten me, what exactly is an mt: rmtio: Input/output error and how can I diagnose where exactly this error is occuring. I have tried disconnecting all other drives on the SCSI bus and only connecting the one scsi drive. I have also tried a different scsi connector on the same cable (which is where another drive that works fine is usually connected). I thought could this be a fault with the cable or the controller? But why would the other 3 drives work fine? Then I thought but if the problem is with my two drives why do they both work in another server? I have also tried connecting one of my working drives to each SCSI connector on the cable in turn and it works fine on every one of them, so surely it is not the SCSI controller or cable? Answers to the obvious: The drive jumper pins are correct I do have a terminator on the end of the SCSI cable I have tried more than one tape I have ran several cleaning cartridges in each drive. There are no error LEDs on the drives I have already tried turning it off and on again :) Please can anyone assist with this? I am completely stumped. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
20.06.2017 19:55, Paul Groves пишет:
Hi,
I have 4 tape drives in my server and I am lately experiencing a few problems.
One night a backup failed with rmtio failed: Input output Error. I ran it again and it worked.
This has happened more and more frequently and now the drive fails writing every time with this error. I put the drive (Hp storageworks Ultrium 920) in another server. It works fine.
I have just received a storageworks ultrium 1760 drive today. Same problem is occuring.
The easiest way to make the error come up is to try mt -f /dev/st3 erase
The drive sounds like it erases the tape, then the tape is unloaded and goes back in again automatically. Then it erases it again, then does it again then again and again then eventually it gives up with the message mt: /dev/st3: rmtioctl failed: Input/output error
Can anyone enlighten me, what exactly is an mt: rmtio: Input/output error and how can I diagnose where exactly this error is occuring.
This is generic catch-all error that simply says that something went wrong. You should have something in kernel log with actual SCSI error seen by tape driver, it may give some hint.
I have tried disconnecting all other drives on the SCSI bus and only connecting the one scsi drive. I have also tried a different scsi connector on the same cable (which is where another drive that works fine is usually connected).
I thought could this be a fault with the cable or the controller? But why would the other 3 drives work fine? Then I thought but if the problem is with my two drives why do they both work in another server?
I have also tried connecting one of my working drives to each SCSI connector on the cable in turn and it works fine on every one of them, so surely it is not the SCSI controller or cable?
Did you try to swap power cables? You did not mention it. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 18:04, Andrei Borzenkov wrote:
This is generic catch-all error that simply says that something went wrong. You should have something in kernel log with actual SCSI error seen by tape driver, it may give some hint. How can I view the kernel log? I have not had to do this on openSUSE before.
Did you try to swap power cables? You did not mention it.
Yes I did. I also tried an external power supply. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/20/2017 09:55 AM, Paul Groves wrote:
Answers to the obvious:
The drive jumper pins are correct I do have a terminator on the end of the SCSI cable I have tried more than one tape I have ran several cleaning cartridges in each drive. There are no error LEDs on the drives I have already tried turning it off and on again :)
Paul.... How sure are you that your terminator is still good. A failing or dodgy terminator can trigger auto-termination built into some drives. (These are often jumper settings to control termination, but some models don't have jumpers but still auto-sense the need for termination, and still others provide no internal termination). These things come to bite you when you move drives after a couple years of use, and forget. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
How sure are you that your terminator is still good. A failing or dodgy terminator can trigger auto-termination built into some drives. When I swap the drive with one of the working ones the same drive has
On 20/06/17 18:14, John Andersen wrote: the fault. Would that prove the terminator is OK?
(These are often jumper settings to control termination, but some models don't have jumpers but still auto-sense the need for termination, and still others provide no internal termination). These things come to bite you when you move drives after a couple years of use, and forget. All of my drives have the same jumper pin layout.
Left to right, Term Pwr / Reserved / SCSIID3 / SCSIID2 / SCSIID1 / SCSIID0 Obviously the 4 right-most ones are the ID number so I have IDs 0, 1, 2 and 3 for my for drives I have no idea what reserved is but none of the drives have a jumper in that position. I also do not know what term pwr (I assume terminator power) is for but all of the drives have this jumper set to on. I assume if I remove the pin from the last drive it will act as a built in terminator? I might be wrong about that but can anyone confirm? All of my drives are HP. one is DAT72 and the other 3 are Storageworks Ultrium drives. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/20/2017 11:43 AM, Paul Groves wrote:
I assume if I remove the pin from the last drive it will act as a built in terminator?
No, that term power just supplies power for the terminator to use to sense if the cable has any devices installed. AFAIR, you only need ONE unit to supply term power and it doesn't matter which one, and having more than one unit supplying term power is perfectly fine. Termination via onboard termination resistors is a different thing. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 19:47, John Andersen wrote:
I assume if I remove the pin from the last drive it will act as a built in terminator? No, that term power just supplies power for the terminator to use to sense if the cable has any devices installed. AFAIR, you only need ONE unit to supply term power and it doesn't matter which one, and having more than one unit supplying term power is
On 06/20/2017 11:43 AM, Paul Groves wrote: perfectly fine.
Termination via onboard termination resistors is a different thing.
You learn something new every day! Thanks for that. I am not sure if the terminator is at fault. I only have the one. If it is the terminator surely it would always be the last drive with the problem? When I connect my LTO2 drive here it works fine. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 20:23, Paul Groves wrote:
On 20/06/17 19:47, John Andersen wrote:
I assume if I remove the pin from the last drive it will act as a built in terminator? No, that term power just supplies power for the terminator to use to sense if the cable has any devices installed. AFAIR, you only need ONE unit to supply term power and it doesn't matter which one, and having more than one unit supplying term power is
On 06/20/2017 11:43 AM, Paul Groves wrote: perfectly fine.
Termination via onboard termination resistors is a different thing.
You learn something new every day! Thanks for that.
I am not sure if the terminator is at fault. I only have the one. If it is the terminator surely it would always be the last drive with the problem? When I connect my LTO2 drive here it works fine.
NO NO NO. My memory of SCSI is patchy, but poor termination messes up the entire cable. Termination is supposed to provide a stable 5V/0V signalling channel or something like that. If the cable isn't properly terminated you get crosstalk, floating voltages, interference - the whole gamut of electrical misbehaviour. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 20:58, Wols Lists wrote:
On 20/06/17 20:23, Paul Groves wrote:
On 20/06/17 19:47, John Andersen wrote:
I assume if I remove the pin from the last drive it will act as a built in terminator? No, that term power just supplies power for the terminator to use to sense if the cable has any devices installed. AFAIR, you only need ONE unit to supply term power and it doesn't matter which one, and having more than one unit supplying term power is
On 06/20/2017 11:43 AM, Paul Groves wrote: perfectly fine.
Termination via onboard termination resistors is a different thing.
You learn something new every day! Thanks for that.
I am not sure if the terminator is at fault. I only have the one. If it is the terminator surely it would always be the last drive with the problem? When I connect my LTO2 drive here it works fine.
NO NO NO.
My memory of SCSI is patchy, but poor termination messes up the entire cable. Termination is supposed to provide a stable 5V/0V signalling channel or something like that. If the cable isn't properly terminated you get crosstalk, floating voltages, interference - the whole gamut of electrical misbehaviour.
Cheers, Wolt OK makes sense to me. I suppose then I should check the voltages on the bus and also take out the cable and check for damage or unusually high resistance on the cores.
Where can anyone even buy SCSI cables and terminators now? Ebay perhaps. It might be a good idea to try a new terminator / cable anyway. I remember SCSI and PATA cables where you crimp your own connectors on. Haven't seen those for a good while now. If it is my cable I might get one of those so I can make the cable as short as possible to avoid unnecessary interference. Saying all this though the other 3 drives operate fine. One thing I noticed is that my SCSI controlled is 160MBps (the same as my DAT72 / LTO1 and LTO2 drives) but the LTO 3 and 4 drives (I am having problems with) are 320MBps. Could it be a compatibility issue perhaps? If I am correct though it should work fine but just at the lower speed. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/20/2017 02:43 PM, Paul Groves wrote:
How sure are you that your terminator is still good. A failing or dodgy terminator can trigger auto-termination built into some drives. When I swap the drive with one of the working ones the same drive has
On 20/06/17 18:14, John Andersen wrote: the fault. Would that prove the terminator is OK?
(These are often jumper settings to control termination, but some models don't have jumpers but still auto-sense the need for termination, and still others provide no internal termination). These things come to bite you when you move drives after a couple years of use, and forget. All of my drives have the same jumper pin layout.
Left to right, Term Pwr / Reserved / SCSIID3 / SCSIID2 / SCSIID1 / SCSIID0
Obviously the 4 right-most ones are the ID number so I have IDs 0, 1, 2 and 3 for my for drives
It's been 15 years since I managed a server, but try changing the ID's to 6,5,4 and 3. I have a vague recollection of higher ID numbers having a higher priority on the SCSI bus. Is this why the controller has number 8? -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/06/17 01:26, Ken Schneider - openSUSE wrote:
It's been 15 years since I managed a server, but try changing the ID's to 6,5,4 and 3. I have a vague recollection of higher ID numbers having a higher priority on the SCSI bus. Is this why the controller has number 8?
Would this really matter seeing as I only ever use one tape drive at a time? Unless you know something I don't? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Tue, Jun 20, 2017 at 2:43 PM, Paul Groves <paul.groves.787@gmail.com> wrote:
Left to right, Term Pwr / Reserved / SCSIID3 / SCSIID2 / SCSIID1 / SCSIID0
Obviously the 4 right-most ones are the ID number so I have IDs 0, 1, 2 and 3 for my for drives
They're binary, so you have addresses 1, 2, 4, and 8. I like to stick to addresses 0-5 if I can. You can do that by pulling off the jumper currently on SCSIID3. That drive will now have address 0, and your other 3 will still have 1, 2, 4. Greg -- Greg Freemyer Advances are made by answering questions. Discoveries are made by questioning answers. — Bernard Haisch -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
I have just run the HP Library and Tape Tools (L&TT) 4.24 on the LTO4 drive and the log is attached below It shows that the tape quick erase is trying to run then the tape is unloaded then loaded. this is then repeating until the drive gives up and fails. What do the errors that have come up mean? Main > Drive Information > Test > Viewer > ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- - Test 'LTO Drive Assessment Test' started on device 'HP Ultrium 4-SCSI' at address 'a0.3.0[10-/dev/sg13]' - Operations Log - executing LTO Drive Assessment Test... - adjusting boost value... - soft unload ... - loading ... - erasing ... - checking tape load ... - erasing ... - soft unload ... - loading ... - writing wrap 42 (1.8 m/sec.) - writing wrap 29 (1.8 m/sec.) - writing wrap 29 (2.1 m/sec.) - writing wrap 29 (2.5 m/sec.) - writing wrap 29 (2.8 m/sec.) - writing wrap 29 (3.1 m/sec.) - writing wrap 29 (3.5 m/sec.) - writing wrap 29 (3.9 m/sec.) - writing wrap 29 (4.2 m/sec.) - writing wrap 29 (1.8 m/sec.) - writing wrap 29 (2.1 m/sec.) - writing wrap 29 (2.5 m/sec.) - writing wrap 29 (2.8 m/sec.) - writing wrap 29 (3.1 m/sec.) - writing wrap 29 (3.5 m/sec.) - writing wrap 29 (3.9 m/sec.) - writing wrap 29 (4.2 m/sec.) - writing wrap 29 (1.8 m/sec.) - writing wrap 29 (2.1 m/sec.) - writing wrap 29 (2.5 m/sec.) - writing wrap 29 (2.8 m/sec.) - soft unload ... - loading ... - erasing ... - Test failed - Analysis Results - LTO Drive Assessment Test, version V11.06.2015 - Test run: Tue Jun 20 18:34:16 2017 - Drive serial number: HU1109FBNH - The Short Erase operation failed to complete. - Sense Key 0x03, Sense Code 0x0c00 (Write error) Error Code: 0x750e miWriteFaultDSPuPVI - Data Cartridge Information: - Vendor: IMATION - Format: LTO-4 - Serial Number: 0D68973290 - Barcode: 000161L4 - There was an unexpected error condition on a Send Diagnostic command. - Byte2 0x3 SenseCode 0xc SenseQual 0x0 (Write error) ErrorCode 0x750e - 1.8 m/sec. tape speed: - Warning (Data written: 1158.3 MB) - 2.1 m/sec. tape speed: - Warning (Data written: 1159.9 MB) - 2.5 m/sec. tape speed: - Warning (Data written: 1159.9 MB) - 2.8 m/sec. tape speed: - Warning (Data written: 1159.9 MB) Enter Command> -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/20/2017 11:47 AM, Paul Groves wrote:
I have just run the HP Library and Tape Tools (L&TT) 4.24 on the LTO4 drive and the log is attached below
- Sense Key 0x03, Sense Code 0x0c00 (Write error) Error Code: 0x750e miWriteFaultDSPuPVI
Paul: Do these tapes have a Use-Cycle management feature that causes them to intentionally fail a life-expectancy test? (Its a feature, not a bug on some drives and some tapes to prevent taking backups that can't be read). https://community.saas.hpe.com/t5/Data-Protector-Practitioners/quot-Cannot-i... No-longer recommended for use. If the same tape works fine in another drive it may be dirty heads in the drive where it fails. It can also indicate the Tape Read or Write Head in the drive has reached end of recommended life. https://community.hpe.com/t5/Tape-Backup-Small-and-Medium/Failed-LTO-drive-a... -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 19:58, John Andersen wrote:
Do these tapes have a Use-Cycle management feature that causes them to intentionally fail a life-expectancy test? As far as I am aware no, the tapes work in other drives fine.
How would I check a tape cartridge for this?
If the same tape works fine in another drive it may be dirty heads in the drive where it fails. Even though I have put several cleaning tapes through?
Does anyone know the procedure to clean the heads manually? I have used alcohol on VHS tape heads before would that work?
It can also indicate the Tape Read or Write Head in the drive has reached end of recommended life. How would I check this? It must record on the drive somewhere surely?
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/20/2017 12:28 PM, Paul Groves wrote:
How would I check this? It must record on the drive somewhere surely?
It might be the result of the self diagnostic tests. If cleaning (per manufacturer directions) doesn't fix it, the heads probably need replacement. Check Vendor's website. Maybe just reach for your checkbook. All my knowledge of tape drives is 10ish years old. I've reached the area of guesswork here. (I fought them long and hard since the Netware 2.x era). Upshot: If you can't trust this drive, you certainly don't want to keep it. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 20:33, John Andersen wrote:
How would I check this? It must record on the drive somewhere surely? It might be the result of the self diagnostic tests. If cleaning (per manufacturer directions) doesn't fix it, the heads
On 06/20/2017 12:28 PM, Paul Groves wrote: probably need replacement. Check Vendor's website. Maybe just reach for your checkbook.
All my knowledge of tape drives is 10ish years old. I've reached the area of guesswork here. (I fought them long and hard since the Netware 2.x era).
Upshot: If you can't trust this drive, you certainly don't want to keep it.
I just found the Device Analysis option and there are no errors on the drives internal log. There must be somewhere to check the estimated life expectancy of the heads. There is an option to test the tape cartridge for wear / data integrity so there must be one for the drive somewhere. I am not sure if I can trust the drive or not yet. it may just be dirty. for all I know the previous owner may never have cleaned the thing. I just read on-line you are supposed to run a cleaning cartridge 5 times to completely clean the heads. I will try this now but the drive is not asking to be cleaned so no idea if this will work. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/06/17 20:33, John Andersen wrote:
If cleaning (per manufacturer directions) doesn't fix it, the heads probably need replacement. Check Vendor's website.
Using HP LTT I found this: It shows that the drive is barely used as it says 2% of lifetime. So I imagine it is dirty heads? - Offtrack margin : Great margin - Channel variation margin : Great margin - Head life : Great margin, 94% life remaining - Reposition life : Great margin, 99% life remaining - Load/unload life : Great margin, 97% life remaining - Drive Environment - Current temperature margin : Great margin (100%) - Maximum temperature margin : Great margin (94%) - Minimum temperature margin : Great margin (100%) - Drive Usage - Drive - Power on time : 1692 days, 12 hours, 51 minutes - Tape Pulled : 978.9 full volume equivalents - Tape Metres Pulled : 42924440 metres - Load/unload cycles : 1,180 (2% of expected life) - Duty cycle : 16% - Power cycles : 18 - Cleans : 8 - Firmware upgrades : 0 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Tue, Jun 20, 2017 at 9:47 PM, Paul Groves <paul.groves.787@gmail.com> wrote: ...
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- - Test 'LTO Drive Assessment Test' started on device 'HP Ultrium 4-SCSI' at ... - Sense Key 0x03, Sense Code 0x0c00 (Write error) Error Code: 0x750e miWriteFaultDSPuPVI
My sources say that 0x750e associated with drive mechanical issues. I have no explanations why it happens in one case but not the other. Have you used the same tape each time? Does it happen on the same tape or on different tapes? Is drive position (i.e. horizontal or vertical) the same in all cases? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/06/17 08:46, Andrei Borzenkov wrote:
My sources say that 0x750e associated with drive mechanical issues. I have no explanations why it happens in one case but not the other. The HP L&TT says there are no mechanical issues. Have you used the same tape each time? Does it happen on the same tape or on different tapes? Is drive position (i.e. horizontal or vertical) the same in all cases?
It is on all tapes regardless of position. LT&T also says I should use a cleaning cartridge, but I have already done this and the clean led on the front is not on. I was reading HP's whitepaper about LTO cleaning (can't remember where I got this) It says if the cleaning LED is off / The error code for cleaning is not present or a cleaning cartridge has been inserted withing 2 weeks it will not clean using the cartridge, instead it will use the internal head cleaning brush. I was wondering, what if this brush is dirty of there is dust / residue / moisture on the heads? Perhaps if I clean them manually would help? Worst case scenario perhaps the heads have corrosion on them? At this point I have no idea what else to do? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Tue, Jun 20, 2017 at 12:55 PM, Paul Groves <paul.groves.787@gmail.com> wrote:
Hi,
I have 4 tape drives in my server and I am lately experiencing a few problems.
One night a backup failed with rmtio failed: Input output Error. I ran it again and it worked.
This has happened more and more frequently and now the drive fails writing every time with this error. I put the drive (Hp storageworks Ultrium 920) in another server. It works fine.
I have just received a storageworks ultrium 1760 drive today. Same problem is occuring.
The easiest way to make the error come up is to try mt -f /dev/st3 erase
The drive sounds like it erases the tape, then the tape is unloaded and goes back in again automatically. Then it erases it again, then does it again then again and again then eventually it gives up with the message mt: /dev/st3: rmtioctl failed: Input/output error
Can anyone enlighten me, what exactly is an mt: rmtio: Input/output error and how can I diagnose where exactly this error is occuring.
I have tried disconnecting all other drives on the SCSI bus and only connecting the one scsi drive. I have also tried a different scsi connector on the same cable (which is where another drive that works fine is usually connected).
I thought could this be a fault with the cable or the controller? But why would the other 3 drives work fine? Then I thought but if the problem is with my two drives why do they both work in another server?
I have also tried connecting one of my working drives to each SCSI connector on the cable in turn and it works fine on every one of them, so surely it is not the SCSI controller or cable?
Answers to the obvious:
The drive jumper pins are correct I do have a terminator on the end of the SCSI cable I have tried more than one tape I have ran several cleaning cartridges in each drive. There are no error LEDs on the drives I have already tried turning it off and on again :)
Please can anyone assist with this? I am completely stumped.
Paul, I've had lots of SCSI issues with tape drives over the years.
From what you describe, I don't think it is the tape drive has failed or needs cleaning because those wouldn't fix themselves by moving from one server to another.
1) Have you tried the drive as the only device on the SCSI bus? I've seen lots of problems go away once the other drives are pulled from the bus. Not always a great solution, but it helps with troubleshooting. 2) I gather you have a physical terminator at the far end of the bus. You haven't talked about the terminator at the other end of the SCSI bus. ie. SCSI requires 2 terminators, one at each end of the bus. Your SCSI controller should have a setting that enables/disables the terminator built into the controller. 3) You've moved the tape drive to another computer and it worked. Have you moved the cable and the drive as a pair? 4) Have you tried a different addresses? Have you reset the address pins and/or swapped them between drives. I've seen address failures in the past. The address pins are binary. You said you have 0, 1, 2, 3. That's 4 bits, so you can pick any address from 0 to 15 if you have 4 jumper pins to work with. Just make sure you don't have an address collision between the controller and the various drives. Good Luck Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
From what you describe, I don't think it is the tape drive has failed or needs cleaning because those wouldn't fix themselves by moving from one server to another. I tried a different SCSI card and cable borrowed from a friend in the server and the problem is still there with both drives. I have since
On 21/06/17 21:51, Greg Freemyer wrote: tried the drives in the other server (where they did work) and now they don't. Perhaps there is some muck in the drive and moving shook it loose?
1) Have you tried the drive as the only device on the SCSI bus? I've seen lots of problems go away once the other drives are pulled from the bus. Not always a great solution, but it helps with troubleshooting. Yes, with both problem drives, no difference in the fault(s) 2) I gather you have a physical terminator at the far end of the bus. You haven't talked about the terminator at the other end of the SCSI bus. ie. SCSI requires 2 terminators, one at each end of the bus. Your SCSI controller should have a setting that enables/disables the terminator built into the controller. I have a physical terminator on the far end. As for the controller end, I never knew this existed but the drives have worked fine for over a year with the current configuration.
How would I check this setting (if the card has this option).
3) You've moved the tape drive to another computer and it worked. Have you moved the cable and the drive as a pair? I didn't at first, then I did and it worked. But as I mentioned above now it does not work in the other server anymore. 4) Have you tried a different addresses? Have you reset the address pins and/or swapped them between drives. I've seen address failures in the past. The address pins are binary. You said you have 0, 1, 2, 3. That's 4 bits, so you can pick any address from 0 to 15 if you have 4 jumper pins to work with. Just make sure you don't have an address collision between the controller and the various drives.
I have checked these. My settings are DAT72 - 0000 LTO1 - 0001 LTO2 - 0010 LTO3 / 4 (when I swap them they have the same ID) - 0011 The SCSI Controlled PCI card is - 0111 (7) which cannot be changed.
Good Luck Greg
I am now suspecting this is not an SCSI problem as with two different cables / servers / SCSI controllers the problem is only the LTO3 and 4 drives, so it must be a problem with the drives. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Paul, The only other thing I can think of is SE / LVD / HVD compatibility (or lack thereof). SE - Single Ended LVD - Low Voltage Differential HVD - High Voltage Differential Especially for the new drive are you sure it is the same as your controller. Basically the controller and everything it talks to should be SE exclusively, or LVD exclusively, or HVD exclusively. It's been a few years since I worked with SCSI routinely, but I don't think there is any compatibility between those varieties at all. You just have to buy all compatible devices. Did you take that into account when you bought the new tape drive? Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/06/17 20:47, Greg Freemyer wrote:
Paul,
The only other thing I can think of is SE / LVD / HVD compatibility (or lack thereof).
SE - Single Ended LVD - Low Voltage Differential HVD - High Voltage Differential
Especially for the new drive are you sure it is the same as your controller. Basically the controller and everything it talks to should be SE exclusively, or LVD exclusively, or HVD exclusively.
It's been a few years since I worked with SCSI routinely, but I don't think there is any compatibility between those varieties at all. You just have to buy all compatible devices.
Did you take that into account when you bought the new tape drive?
Greg
Yes I did. My controller and all tape drives are all LVDS -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
The seller has taken my LTO4 tape drive back and is going to get an engineer to check the heads. Will update when I hear back. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
So My LTO4 drive had dodgy tape heads. I have been refunded and managed to obtain another couple of HP Storageworks 1760 drives. I am still experiencing odd problems though. These are my drives: HP DAT72] HP Storageworks 232 LTO1 HP Storageworks 448 LTO2 HP Storageworks 920 LTO3 HP Storageworks 1760 LTO4 x2 All of them are SCSI LVDS drives. The DAT72 is 40mbps the LTO1 and 2 are 160mbps and the LTO3 and 4 are 320mbps. If I connect the DAT72 / LTO1 / LTO2 all 3 work fine. If I add in the LTO3 or either LTO4 drive, the LTO3 or 4 drives do not work (They start writing then suddenly stop dead and rewind the tape and I cannot terminate the process, tar / mt or HP L&TT, even with kill -9 unless I reboot). If I connect up one LTO4 or just the LTO3 on its own, with no other drives on the SCSI bus, then they work fine and pass all the tests in the HP L&TT program. So I would think that the drives themselves are OK because they all appear to work fine individually. If I connect up an LTO4 drive and the LTO3 drive both do not work (same problem as before). If I connect an LTO4 drive with the LTO2 drive, the LTO4 drive will not work (same problem) but the LTO2 does. The same happens if I connect the LTO3 drive with the LTO2 drive, the LTO3 drive will not work (same problem) but the LTO2 does. The weird thing is that for over a year I used my lto3 drive at the same time as the lto2 drive with no problems. If I connect both LTO4 drives they do not work (again, same problem). I have replaced the controller card and SCSI cable and also the terminator but none of these made a difference to the problem. I have also checked the power supply and all of the voltages are as expected and it is only at 12% load. Anyone have any ideas what is happening? Thanks Paul -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/08/2017 11:16 AM, Paul Groves wrote:
Anyone have any ideas what is happening?
Jeeze, you've been through the wringer on this one... About here is where I scream "Bad Capacitors" and run away like my hair is on fire. (Bad caps often meter just fine, until you put a load on the power supply). You's have to open that power supply to see them. Got Spare? Swap it in just to cover all bases. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/08/2017 02:16 PM, Paul Groves wrote:
So My LTO4 drive had dodgy tape heads. I have been refunded and managed to obtain another couple of HP Storageworks 1760 drives.
I am still experiencing odd problems though.
These are my drives:
HP DAT72] HP Storageworks 232 LTO1 HP Storageworks 448 LTO2 HP Storageworks 920 LTO3 HP Storageworks 1760 LTO4 x2
All of them are SCSI LVDS drives. The DAT72 is 40mbps the LTO1 and 2 are 160mbps and the LTO3 and 4 are 320mbps.
If I connect the DAT72 / LTO1 / LTO2 all 3 work fine. If I add in the LTO3 or either LTO4 drive, the LTO3 or 4 drives do not work (They start writing then suddenly stop dead and rewind the tape and I cannot terminate the process, tar / mt or HP L&TT, even with kill -9 unless I reboot).
If I connect up one LTO4 or just the LTO3 on its own, with no other drives on the SCSI bus, then they work fine and pass all the tests in the HP L&TT program.
So I would think that the drives themselves are OK because they all appear to work fine individually.
If I connect up an LTO4 drive and the LTO3 drive both do not work (same problem as before).
If I connect an LTO4 drive with the LTO2 drive, the LTO4 drive will not work (same problem) but the LTO2 does.
The same happens if I connect the LTO3 drive with the LTO2 drive, the LTO3 drive will not work (same problem) but the LTO2 does. The weird thing is that for over a year I used my lto3 drive at the same time as the lto2 drive with no problems.
If I connect both LTO4 drives they do not work (again, same problem).
I have replaced the controller card and SCSI cable and also the terminator but none of these made a difference to the problem. I have also checked the power supply and all of the voltages are as expected and it is only at 12% load.
Anyone have any ideas what is happening?
Thanks
Paul
Is each device using it's own termination? Only the last device should terminate the SCSI bus. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 09/07/17 14:52, Ken Schneider - openSUSE wrote:
Is each device using it's own termination? Only the last device should terminate the SCSI bus.
No there is only one terminator. My configuration is like this: PCI Card Drive 1 Drive 2 Drive 3 Drive4 Terminator -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 09/07/17 15:37, Paul Groves wrote:
On 09/07/17 14:52, Ken Schneider - openSUSE wrote:
Is each device using it's own termination? Only the last device should terminate the SCSI bus.
No there is only one terminator.
My configuration is like this:
PCI Card Drive 1 Drive 2 Drive 3 Drive4 Terminator
No that's not the question. A lot of SCSI devices come with built-in termination. Check the jumper documentation, and make sure all the internal terminators are disabled if you've got an external terminator. If you have devices on all the blocks, there's nowhere to put a terminator which is why they do this, and which is why you should always have the terminator - or a terminating device - on the last block. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 09/07/17 17:23, Wols Lists wrote:
No that's not the question. A lot of SCSI devices come with built-in termination. Check the jumper documentation, and make sure all the internal terminators are disabled if you've got an external terminator.
If you have devices on all the blocks, there's nowhere to put a terminator which is why they do this, and which is why you should always have the terminator - or a terminating device - on the last block.
Cheers, Wol I can't see any mention in the manual of built in terminators. The jumper pins on all the drives are
Term Pwr Reserved SCSI_ID 4 SCSI_ID 3 SCSI_ID 2 SCSI_ID 1 I assume the Terminator power supplies the power to the terminator on the last block? This jumper is set to on on all the drives. I have no idea what the reserved jumper is but there is no jumper connector provided for this at all on any of the drives -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/09/2017 01:16 PM, Paul Groves wrote:
Term Pwr Reserved SCSI_ID 4 SCSI_ID 3 SCSI_ID 2 SCSI_ID 1
Is there any potential for ID conflict with 3 & 4? I don't do SCSI so that may not even apply... -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sun, Jul 9, 2017 at 3:20 PM, David C. Rankin <drankinatty@suddenlinkmail.com> wrote:
On 07/09/2017 01:16 PM, Paul Groves wrote:
Term Pwr Reserved SCSI_ID 4 SCSI_ID 3 SCSI_ID 2 SCSI_ID 1
Is there any potential for ID conflict with 3 & 4? I don't do SCSI so that may not even apply...
Those are bits. ID can be 0-15 and have to be unique per device. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
I have tried setting different IDs and it doesn't appear to make a difference. Currently they're set to: 0001 0100 1000 1001 On 09/07/17 20:20, David C. Rankin wrote:
On 07/09/2017 01:16 PM, Paul Groves wrote:
Term Pwr Reserved SCSI_ID 4 SCSI_ID 3 SCSI_ID 2 SCSI_ID 1
Is there any potential for ID conflict with 3 & 4? I don't do SCSI so that may not even apply...
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/09/2017 04:44 PM, Paul Groves wrote:
I have tried setting different IDs and it doesn't appear to make a difference. Currently they're set to:
0001
0100
1000
1001
Are those jumpers (e.g. two jumpers on last one) or bits? (e.g. 1, 4, 8 & 9) -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 09/07/17 23:00, David C. Rankin wrote:
On 07/09/2017 04:44 PM, Paul Groves wrote:
I have tried setting different IDs and it doesn't appear to make a difference. Currently they're set to:
0001
0100
1000
1001 Are those jumpers (e.g. two jumpers on last one) or bits? (e.g. 1, 4, 8 & 9)
They are bits so it is drive 1, 4, 8 and 9 (controller card is set to 7). I have also tried various other combinations of ID numbers on the drives but this does not seem to have affected the problem. I used to have it quite nicely set up before it went wrong: 0000 - /dev/st0 - DAT72 0001 - /dev/st1 - LTO1 0010 - /dev/st2 - LTO2 0011 - /dev/st3 - LTO3 That is until I decided to replace the LTO3 drive for the LTO4 now all hell has broken loose! I think I will next try physically moving the drives about in order on the scsi bus. I doubt this will help in any way as I do not see how the drives know what order they are physically connected, only their ID number set by the jumper pins. But who knows, maybe some electrical interference between drives? Or compatibility (even though they are all hp drives? Would it be a good idea to upgrade the drive firmware. They are all asking me to in LTT but I never did it because they worked fine. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-07-10 13:10, Paul Groves wrote:
I think I will next try physically moving the drives about in order on the scsi bus. I doubt this will help in any way as I do not see how the drives know what order they are physically connected, only their ID number set by the jumper pins.
But who knows, maybe some electrical interference between drives? Or compatibility (even though they are all hp drives?
On SCSSI it does matter. Similarly, you need to place terminator resistors at the end (both ends) of the cable run - or you get signal bounces and interferences. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 10/07/17 14:19, Carlos E. R. wrote:
On 2017-07-10 13:10, Paul Groves wrote:
I think I will next try physically moving the drives about in order on the scsi bus. I doubt this will help in any way as I do not see how the drives know what order they are physically connected, only their ID number set by the jumper pins.
But who knows, maybe some electrical interference between drives? Or compatibility (even though they are all hp drives? On SCSSI it does matter. Similarly, you need to place terminator resistors at the end (both ends) of the cable run - or you get signal bounces and interferences.
Would my SCSI Card have terminator resistors built in? I assume it must do because it has worked for over a year and a half with no problems and works with my other drives. There are four jumpers on my PCI card, two labelled LED (obviously just the indicator so unimportant seeing as the drives have their own LEDs) and two with no label. Would these have something to do with it? There is no pin connecting them. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/10/2017 08:25 AM, Paul Groves wrote:
Would my SCSI Card have terminator resistors built in?
Look for physical terminators or dual cable sockets. There is no rule that the card must be at the end of a cable. SCSI cards can be in the middle, with devices on both sides. Its unusual to find a single jumper controlling the resistors on 8, 16, or 24 parallel lines. It costs more electronics to do that than just include a 12 cent resistor pack. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, Jul 10, 2017 at 12:51 PM, John Andersen <jsamyth@gmail.com> wrote:
On 07/10/2017 08:25 AM, Paul Groves wrote:
Would my SCSI Card have terminator resistors built in?
Look for physical terminators or dual cable sockets.
There is no rule that the card must be at the end of a cable. SCSI cards can be in the middle, with devices on both sides.
Its unusual to find a single jumper controlling the resistors on 8, 16, or 24 parallel lines. It costs more electronics to do that than just include a 12 cent resistor pack.
I don't have SCSI in my systems anymore, but the last controllers I used had a BIOS setting to enable/disable the controller termination resistors. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-07-10 20:51, Greg Freemyer wrote:
On Mon, Jul 10, 2017 at 12:51 PM, John Andersen <jsamyth@gmail.com> wrote:
On 07/10/2017 08:25 AM, Paul Groves wrote:
Would my SCSI Card have terminator resistors built in?
Look for physical terminators or dual cable sockets.
There is no rule that the card must be at the end of a cable. SCSI cards can be in the middle, with devices on both sides.
Its unusual to find a single jumper controlling the resistors on 8, 16, or 24 parallel lines. It costs more electronics to do that than just include a 12 cent resistor pack.
I don't have SCSI in my systems anymore, but the last controllers I used had a BIOS setting to enable/disable the controller termination resistors.
The resistor block is very cheap itself, but it uses board space plus a connector and strong soldering. Board space is paid by square centimetre at manufacturing, and is also limited otherwise. Thus it maybe cheaper to use electronics. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 10/07/17 19:51, Greg Freemyer wrote:
I don't have SCSI in my systems anymore, but the last controllers I used had a BIOS setting to enable/disable the controller termination resistors.
Greg Same here :) In the card's BIOS
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-07-10 17:25, Paul Groves wrote:
On 10/07/17 14:19, Carlos E. R. wrote:
On 2017-07-10 13:10, Paul Groves wrote:
I think I will next try physically moving the drives about in order on the scsi bus. I doubt this will help in any way as I do not see how the drives know what order they are physically connected, only their ID number set by the jumper pins.
But who knows, maybe some electrical interference between drives? Or compatibility (even though they are all hp drives? On SCSSI it does matter. Similarly, you need to place terminator resistors at the end (both ends) of the cable run - or you get signal bounces and interferences.
Would my SCSI Card have terminator resistors built in? I assume it must do because it has worked for over a year and a half with no problems and works with my other drives.
Well, look at the photos John Andersen posted in those links. I don't have any SCSSI hardware, but I guess that a low voltage resistor tester would find out. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On Mon, 10 Jul 2017 16:25:41 +0100 Paul Groves <paul.groves.787@gmail.com> wrote:
On 10/07/17 14:19, Carlos E. R. wrote:
On 2017-07-10 13:10, Paul Groves wrote:
I think I will next try physically moving the drives about in order on the scsi bus. I doubt this will help in any way as I do not see how the drives know what order they are physically connected, only their ID number set by the jumper pins.
But who knows, maybe some electrical interference between drives? Or compatibility (even though they are all hp drives? On SCSSI it does matter. Similarly, you need to place terminator resistors at the end (both ends) of the cable run - or you get signal bounces and interferences.
Would my SCSI Card have terminator resistors built in? I assume it must do because it has worked for over a year and a half with no problems and works with my other drives.
There are four jumpers on my PCI card, two labelled LED (obviously just the indicator so unimportant seeing as the drives have their own LEDs) and two with no label. Would these have something to do with it? There is no pin connecting them.
It's late and I'm jumping in without reading the whole thread, and it's a long time since I did anything with SCSI but ... Has Paul actually posted full details of all his equipment anywhere? Exactly what adapter and drives and cables and how plugged together? Are these drives internal or external and mounted on which side of the controller/adapter, for example? Has he read and understood all the relevant manuals or does he still have questions? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
OK can anyone work this one out? If I connect my drives in this order they now appear to work fine. If I swap any of them over to any other combination I get the random read-write errors I was getting before. SCSI Card (Termination set to on) HP DAT72 HP LTO2 HP LTO1 HP LTO4 Terminator (built onto cable) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/07/17 15:50, Paul Groves wrote:
OK can anyone work this one out?
If I connect my drives in this order they now appear to work fine. If I swap any of them over to any other combination I get the random read-write errors I was getting before.
SCSI Card (Termination set to on) HP DAT72 HP LTO2 HP LTO1 HP LTO4 Terminator (built onto cable) I'm back again! :)
OK, so after 2 successful backups on my LTO4 drive, I set another one off yesterday and lo-and-behold it has broken again! It is doing the same thing as before, I set off the writing to tape whether it be using tar or just dd/catt-ing 1s and 0s to the tape, the drive. It starts to write to the tape then stops dead. Most of the time with the error rmtio Error: failed to write filemark. What exactly does this error mean? If I erase the tape it works fine and if I run the HP LTT Read write test it works fine. So as before, I thought one of the other drives might be playing it up. I disconnected all other drives, problem still occurs. (First time this has happened). I replaced the SCSI controller with a different one. Problem still there. Also changed the SCSI cable again. Problem still occurs. This is happening for two different HP 1760 and also my LTO3 Ultrium 920 drive. The only common thing here is that these 3 drives are 320Mbps SCSI LVD and all my others are 160 or 40 mbps LVD. Could this be a problem? My cards are all 160MBps. Saying that though the LTO3 tape drive worked fine for ages (over a year) with no trouble on the same hardware. I have not installed any updates since before I had the problems and nothing has been changed configuration-wise with the server other than removing the un-neccesary graphics card (re-installing it doesn't affect the problem). Any advice on how can I diagnose this problem further? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 7/18/2017 1:56 PM, Paul Groves wrote:
Any advice on how can I diagnose this problem further?
Your controllers send data slower, so that should make for less errors. I can't believe that is the problem. Failed to write tape mark happens at the start and end of each backup. These demark the backup sessions. I remember something about writing a tape mark, backing up over it, and then going forward again reading the tapemark. If it can't find the tapemark, it assumes it was not able to write it. Its really that it couldn't read it. Things that caused that (back in the day when I was using tape) were: tape leaders of wrong type/lenght (or leaders when the drive didn't want any, worn out first portion of the tape (it gets more wear), improper erase, tape mark from an incompitable drive unexpectedly encountered. (in those days we really couldn't use the same tapes on different models of drives - maybe that is fixed these days). Cheapo off brand tapes, -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 19/07/17 03:59, John M Andersen wrote:
Your controllers send data slower, so that should make for less errors. I can't believe that is the problem. I would have thought this too. Failed to write tape mark happens at the start and end of each backup. These demark the backup sessions. It said filemark not tape mark. I remember something about writing a tape mark, backing up over it, and then going forward again reading the tapemark. If it can't find the tapemark, it assumes it was not able to write it. Its really that it couldn't read it. How would I prove this? Things that caused that (back in the day when I was using tape) were: tape leaders of wrong type/lenght (or leaders when the drive didn't want any, I thought Ultrium LTO were all the same standard? The tapes are being wound onto the take-up reel in the drive just fine. I can see it goind round through the air-vent in the back of the drive. worn out first portion of the tape (it gets more wear), They are second hand tapes but all of them have the problem. In HP LTT it says the cartridges are hardly used. improper erase, I used mt -f /dev/st* erase 0 (for quick erase) Which I always do before a backup. tape mark from an incompitable drive unexpectedly encountered. (in those days we really couldn't use the same tapes on different models of drives - maybe that is fixed these days). It is a HP drive, would that be a problem. Cheapo off brand tapes, I have tried with Sony, TDK, Imation and IBM LTO4 tape cartridges. I don't have any HP ones but I do have some HP LTO3 I haven't tried yet. I will try it with those.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
I am completely at odd with this problem. I have connected up both LTO 4 tape drives one at a time to the server and removed all other unnecessary hardware and I am still getting the original problem. Which is; It starts writing to the tape and then the drive makes a knocking sound once and stops. Leaving the command frozen and unable to cancel with Ctrl + C, Ctrl + Z or Ctrl + \ leaving me with no option other than to reboot. Same results with the LTO3 drive. Again, all my other tape drives work fine, the only common thing between the drives that are not working that I can see is that the drives are 320Mbps LVD not 160 LVD like the controller and other drives. I am pretty much stumped. It seems highly unlikely that the 4 different drives are all faulty with exactly the same problem. I have ordered a new SCSI cable and a 320Mbps PCI card. I will try with just the one drive with these and see what happens. Unless anyone else can suggest anything? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/21/2017 10:51 AM, Paul Groves wrote:
I have ordered a new SCSI cable and a 320Mbps PCI card. I will try with just the one drive with these and see what happens.
Unless anyone else can suggest anything?
That sounds like a logical approach. You may end up finding some inherent incompatibility between the use of that combination of 4 drives at once, or it may come down to a bad cap in one of them that only comes into play when all 4 are on the same cable. I don't know what your backup is like (whether you write to all 4 simultaneously or if it is a serial backup to different drives) In the 2nd case, you can always split the drive between 2 machines and see if that is a way around the issue. I feel your pain. There are some things in this world you just can't explain (there was a song or something like that once...) -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/21/2017 08:51 AM, Paul Groves wrote:
I have ordered a new SCSI cable and a 320Mbps PCI card.
Well you've done everything else. Are we sure that 320 Mbps cards are compatible with 160 Mbps drives? I've found one page (somewhat dated) that suggests there is quite a bit of difference between the two standards. https://kb.iu.edu/d/aiqw Maybe you will have to keep both cards, and segregate devices accordingly. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/07/17 18:36, John Andersen wrote:
On 07/21/2017 08:51 AM, Paul Groves wrote:
I have ordered a new SCSI cable and a 320Mbps PCI card. Well you've done everything else.
Are we sure that 320 Mbps cards are compatible with 160 Mbps drives? I've found one page (somewhat dated) that suggests there is quite a bit of difference between the two standards. https://kb.iu.edu/d/aiqw
Maybe you will have to keep both cards, and segregate devices accordingly.
I bet everyone if fed up with this by now. :) So, my Adaptec Ultra 320 card has arrived. I connected up the card to the server with only the HP 1760 LTO4 drive connected. No other SCSI devices. I turned on termination in the controllers bios screen and connected the terminator on the end of the cable after the drive. At boot it detects it as a Ultra 320 device and there are no errors on the controller side. lsscsi shows the drive fine. I ran mt -f /dev/st0 erase 0 the drive started to erase, at the point it started writing to the tape the same problem occurs. The problem: The drive goes 'donk' and stops dead. The tape spools for some time back and forth and repeats occasionally 'donk'ing. Eventually the drive gives up and ejects the tape then takes it back in. This repeats forever until I eject the cartridge. The server at this point is frozen on the command still. I cannot Ctrl + C, Ctrl + Z or Ctrl + \ I have no choice but to force reboot using sysrq keys. The same has happened with both of my LTO4 drives. Both HP Storageworks 1760 At this point I am stumped. I have changed the cable, the controller and the drive and it still is not working. :( -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, Aug 5, 2017 at 8:41 AM, Paul Groves <paul.groves.787@gmail.com> wrote:
On 21/07/17 18:36, John Andersen wrote:
On 07/21/2017 08:51 AM, Paul Groves wrote:
I have ordered a new SCSI cable and a 320Mbps PCI card.
Well you've done everything else.
Are we sure that 320 Mbps cards are compatible with 160 Mbps drives? I've found one page (somewhat dated) that suggests there is quite a bit of difference between the two standards. https://kb.iu.edu/d/aiqw
Maybe you will have to keep both cards, and segregate devices accordingly.
I bet everyone if fed up with this by now. :)
So, my Adaptec Ultra 320 card has arrived. I connected up the card to the server with only the HP 1760 LTO4 drive connected. No other SCSI devices.
I turned on termination in the controllers bios screen and connected the terminator on the end of the cable after the drive.
At boot it detects it as a Ultra 320 device and there are no errors on the controller side.
lsscsi shows the drive fine.
I ran mt -f /dev/st0 erase 0
the drive started to erase, at the point it started writing to the tape the same problem occurs.
The problem:
The drive goes 'donk' and stops dead. The tape spools for some time back and forth and repeats occasionally 'donk'ing. Eventually the drive gives up and ejects the tape then takes it back in. This repeats forever until I eject the cartridge.
The server at this point is frozen on the command still. I cannot Ctrl + C, Ctrl + Z or Ctrl + \
I have no choice but to force reboot using sysrq keys.
The same has happened with both of my LTO4 drives. Both HP Storageworks 1760
At this point I am stumped. I have changed the cable, the controller and the drive and it still is not working. :(
Paul, I've never had had that much trouble with SCSI, but one time I was installing a system in the south of France. It just wasn't reliable. The solution: go outside and drive a 20-foot stake in the ground, then connect the building ground to it. So, can you transport your system to somewhere else that likely has good quality power and test it there? If that fixes it, you have a power/ground problem. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/05/2017 08:24 AM, Greg Freemyer wrote:
At this point I am stumped. I have changed the cable, the controller and the drive and it still is not working. :( Paul,
I've never had had that much trouble with SCSI, but one time I was installing a system in the south of France. It just wasn't reliable.
The solution: go outside and drive a 20-foot stake in the ground, then connect the building ground to it.
---edit--- The solution: go outside and 'place your computer on the ground, place a 20-foot stake on the mid-point of your case laid horizontally and drive the 20-foot stake into the ground'. ------- On a more serious note, I would look to turn any logging or debug associated with any of the tape applications you are using (if it is just tar, you can look at some of the --ignore-command-error or --format options) There is some bit that is being set wrong somewhere, whether it is a result of a drive failure, format problem or what, I'm out of guesses :( -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/08/17 05:59, David C. Rankin wrote:
The solution: go outside and 'place your computer on the ground, place a 20-foot stake on the mid-point of your case laid horizontally and drive the 20-foot stake into the ground'. Please don't tempt me :) On a more serious note, I would look to turn any logging or debug associated with any of the tape applications you are using (if it is just tar, you can look at some of the --ignore-command-error or --format options) I am just using tar, but it also happens when I run the drive assessment test in HP LT. I haven't tried the commands yet. There is some bit that is being set wrong somewhere, whether it is a result of a drive failure, format problem or what, I'm out of guesses:( I have noticed that since installing the Ultra 320 card (adaptec) I have some slightly higher detail in HP Library & Tape Tools Health report.
It has logged the following two messages every time I experience the problem: ISCSI TaskMgmtFunction invoked: SI_LOGICAL_UNIT_RESET_MGMT_FUNC, Initiator ID: 0000h, Via SI_PRIMARY_PORT_ID, To LUN: 0h BusRst ISCSI Bus Reset Occured I am guessing this is my problem. Now what does all of this mean? Does anyone know how I can diagnose further with this information? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/07/2017 12:45 PM, Paul Groves wrote:
It has logged the following two messages every time I experience the problem:
ISCSI TaskMgmtFunction invoked: SI_LOGICAL_UNIT_RESET_MGMT_FUNC, Initiator ID: 0000h, Via SI_PRIMARY_PORT_ID, To LUN: 0h
BusRst ISCSI Bus Reset Occured
I am guessing this is my problem. Now what does all of this mean?
Does anyone know how I can diagnose further with this information?
Progress can be painfully slow, but that IS progress.... I don't have any off the top of the head knowledge of the various defined constant errors themselves, but it does give fresh search terms to drive yourself batty trying various phrases of.... Keep the faith, you'll get there... -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/08/17 03:30, David C. Rankin wrote:
On 08/07/2017 12:45 PM, Paul Groves wrote:
It has logged the following two messages every time I experience the problem:
ISCSI TaskMgmtFunction invoked: SI_LOGICAL_UNIT_RESET_MGMT_FUNC, Initiator ID: 0000h, Via SI_PRIMARY_PORT_ID, To LUN: 0h
BusRst ISCSI Bus Reset Occured
I am guessing this is my problem. Now what does all of this mean?
Does anyone know how I can diagnose further with this information? Progress can be painfully slow, but that IS progress....
I don't have any off the top of the head knowledge of the various defined constant errors themselves, but it does give fresh search terms to drive yourself batty trying various phrases of....
Keep the faith, you'll get there...
This is very slowly driving me to insanity. Does anyone know exactly what this means / how it may be caused? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/08/17 14:24, Greg Freemyer wrote:
Paul,
I've never had had that much trouble with SCSI, but one time I was installing a system in the south of France. It just wasn't reliable.
The solution: go outside and drive a 20-foot stake in the ground, then connect the building ground to it.
So, can you transport your system to somewhere else that likely has good quality power and test it there? If that fixes it, you have a power/ground problem.
Greg
Hi Greg. Interesting idea that it could be an earth problem. I performed an earth continuity test on the server between the drive's casing and the earth in the power cable and this is showing up as all OK. I have also been zipping round plug sockets checking for earth leakage. I cannot find anything above 0.002mV which is absolutely fine. I have also fully PAT Tested my server and the UPS it is connected to and these both pass fine on all tests (Including earth continuity). I have also checked the earth live and neutral terminals for loose screws all the way back to the consumer unit. These are all OK and are making goof connections. So unless something is causing some other type of interference on the earth? Out earth connection is via the Neutral cable coming into the building. We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.). Perhaps this could be causing the problem? I have yet to transport the system to a different site to test it. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-07 19:43, Paul Groves wrote:
So unless something is causing some other type of interference on the earth? Out earth connection is via the Neutral cable coming into the building.
I don't understand this sentence. You don't have a separate ground wire, instead you use the neutral cable? The building doesn't have it's own ground pole, a metal rod inserted into the earth? I'm not familiar with the British electricity code. I guess it will be different! :-p
We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.).
They need to measure it themselves with their own equipment, and I suppose they are very hard to convince to do the measurement.
Perhaps this could be causing the problem?
Dunno.
I have yet to transport the system to a different site to test it.
If that worked, it would nail the problem, certainly. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/07/2017 02:02 PM, Carlos E. R. wrote:
I'm not familiar with the British electricity code. I guess it will be different!
Yes, start with that. (The not familiar part). There are a ton of differences in EU and North American power systems as far as physical structure. Its easy to be less familiar than you think you are. This stuff has become more than a little complex with the growth of grids. Neutral is tied to Ground almost everywhere in the US, if not on your house then on the next nearest pole or power pedestal. Ground rods are not universally used, because the rely on soil moisture content and can behave very inconsistently. During the dry season its common for some large networks to lose their ground because they relied too heavily on grounding rods which were no longer in good contact. Then there are differences between three wire systems (pos, neg, neut) and 4 wire systems (pos net neut ground) that occur in different parts of the world. Some parts of the US have individual houses all on one side (phase), and alternate every other house to the other leg to balance the phases. Then someone strings an extension cord. (Indications are that the US may slowly be moving toward a 4 wire infrastructure. https://en.wikipedia.org/wiki/Three-phase_electric_power But as far as These Tape Drives, this seems unlikely to be a ground issue unless some parts of the tape backup system are on the opposite leg. Electrical code frowns on this and wants all outlets in the same room to be on the same leg. There could also be large ground differences if some parts of the tape system are on a UPS and the other parts are Not, or are on a different UPS. A multi-meter can usually detect this situation for by checking for voltage between ground sockets between outlets. -- After all is said and done, more is said than done.
On 08/07/2017 06:02 PM, John Andersen wrote:
There are a ton of differences in EU and North American power systems as far as physical structure. Its easy to be less familiar than you think you are. This stuff has become more than a little complex with the growth of grids.
Neutral is tied to Ground almost everywhere in the US, if not on your house then on the next nearest pole or power pedestal.
Ummm... Neutral is *ALWAYS* tied to ground at the power entrance. Other services are likewise bonded to the same ground point.
Ground rods are not universally used, because the rely on soil moisture content and can behave very inconsistently. During the dry season its common for some large networks to lose their ground because they relied too heavily on grounding rods which were no longer in good contact.
Actually, ground rods are used, as water mains are no longer a reliable ground. There should also be a ground wire running back to the power line.
Then there are differences between three wire systems (pos, neg, neut) and 4 wire systems (pos net neut ground) that occur in different parts of the world. Some parts of the US have individual houses all on one side (phase), and alternate every other house to the other leg to balance the phases. Then someone strings an extension cord. (Indications are that the US may slowly be moving toward a 4 wire infrastructure. https://en.wikipedia.org/wiki/Three-phase_electric_power
There's no positive or negative with AC power. It's constantly changing polarity. There are 1 or more hot wires and neutral, with ground completely isolated, except where it connects to the neutral at the power entrance. In normal house wiring there are 2 hot wires, 180° apart, with most circuits using only one hot. Heavy loads, such as stoves & dryers, use both. Three phase has been around since Tesla and is used in industrial or large commercial installations. I doubt we'll see 3 phase in homes any time soon, as it would require a lot of major changes, including replacing any 240V appliances. Incidentally, 3 phase only uses 4 wires with the wye configuration, where there is a common point that can be grounded. With the delta configuration, there is no common point that can be grounded and so no 4th wire. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/07/2017 03:21 PM, James Knott wrote:
John Andersen wrote
Neutral is tied to Ground almost everywhere in the US, if not on your house then on the next nearest pole or power pedestal.
Ummm... Neutral is *ALWAYS* tied to ground at the power entrance. Other services are likewise bonded to the same ground point.
We said the same thing, BUT, I was REALLY pointing out (somewhat off topic) is that it is not uncommon in certain geological regions for there to be a difference between grounded neutral and actual ground simply because of a loss of ground-conductivity. In some systems (cable tv systems often), this leads to the entire cable plant floating above ground by a measurable (but not dangerous) volt or two, which plays havoc with reception. This was common in some parts of alaska, probably due to inconsistant grounding practices of the installers and the rush job they did when installing. But more to the point there are differences in grounding practices in some parts of the world: https://en.wikipedia.org/wiki/Ground_and_neutral#Grounding_systems TN-C and TN-S. Quoting above....
In the TN-C system, a common conductor provides both the neutral and protective grounding. The neutral conductor is connected to earth ground at the point of supply, and equipment cases are connected to the neutral. The danger exists that a broken neutral connection will allow all the equipment cases to rise to a dangerous voltage if any leakage or insulation fault exists in any equipment. This can be mitigated with special cables but the cost is then higher.
https://en.wikipedia.org/wiki/Earthing_system#IEC_terminology Usually you won't have any problem in the same data center. But as common office space get pressed into service as mini-data-centers, all bets are off. **In any event, if all the SCSI devices are plugged into the same wall socket or UPS there should not be any grounding issues.** There's a lot of mythology involved here as well: http://www.eetimes.com/document.asp?doc_id=1272972 -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 01:07, John Andersen wrote:
But more to the point there are differences in grounding practices in some parts of the world:
Yes.
https://en.wikipedia.org/wiki/Ground_and_neutral#Grounding_systems TN-C and TN-S.
That's what I was looking for, thanks. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/07/2017 07:07 PM, John Andersen wrote:
There's a lot of mythology involved here as well: http://www.eetimes.com/document.asp?doc_id=1272972
I found one error in that article: "In North America, minicomputers frequently are designed to use a 208V 3-wire system. In this 3-wire system there are 2 hot wires and 1 ground wire (no neutral wire)." 208V is the voltage between any 2 phases of a 3 phase circuit, where each phase is 120V to neutral. I suspect they meant 240V. I have worked with a lot of minicomputers and the power supplies I saw were either 120V or 240V, single phase. However, the disk pack drives of that era often had 208V 3 phase power. BTW, the tape drives were 120V, to keep on topic. ;-) In the North American power system, local power feeds are from a centre tapped transformer that's connected across 1 of the 3 phases of a higher voltage line. This results in a 120/240V single phase connection to a home. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/07/2017 07:07 PM, John Andersen wrote:
But more to the point there are differences in grounding practices in some parts of the world: https://en.wikipedia.org/wiki/Ground_and_neutral#Grounding_systems TN-C and TN-S.
Quoting above....
In the TN-C system, a common conductor provides both the neutral and protective grounding. The neutral conductor is connected to earth ground at the point of supply, and equipment cases are connected to the neutral. The danger exists that a broken neutral connection will allow all the equipment cases to rise to a dangerous voltage if any leakage or insulation fault exists in any equipment. This can be mitigated with special cables but the cost is then higher. https://en.wikipedia.org/wiki/Earthing_system#IEC_terminology
North America would be TN-C-S, with separate neutral and ground wires, joined only at the power entrance inside, but the power is delivered to the building with a ground wire and hot wires, but no separate neutral. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 00:21, James Knott wrote:
On 08/07/2017 06:02 PM, John Andersen wrote:
There are a ton of differences in EU and North American power systems as far as physical structure. Its easy to be less familiar than you think you are. This stuff has become more than a little complex with the growth of grids.
Neutral is tied to Ground almost everywhere in the US, if not on your house then on the next nearest pole or power pedestal.
Ummm... Neutral is *ALWAYS* tied to ground at the power entrance. Other services are likewise bonded to the same ground point.
Not in Spain, I'm certain of that. It is connected, I understand, at the step down transformer, which can serve several city blocks. I don't know in UK, which is the point of interest for the OP.
There's no positive or negative with AC power. It's constantly changing polarity. There are 1 or more hot wires and neutral, with ground completely isolated, except where it connects to the neutral at the power entrance. In normal house wiring there are 2 hot wires, 180° apart, with most circuits using only one hot.
That's American usage, but not EU. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/07/2017 07:47 PM, Carlos E. R. wrote:
Ummm... Neutral is *ALWAYS* tied to ground at the power entrance. Other
services are likewise bonded to the same ground point. Not in Spain, I'm certain of that.
It is connected, I understand, at the step down transformer, which can serve several city blocks.
That sounds dangerous. It could result in significant voltage differences between the powered device and local ground. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 03:22, James Knott wrote:
On 08/07/2017 07:47 PM, Carlos E. R. wrote:
Ummm... Neutral is *ALWAYS* tied to ground at the power entrance. Other
services are likewise bonded to the same ground point. Not in Spain, I'm certain of that.
It is connected, I understand, at the step down transformer, which can serve several city blocks.
That sounds dangerous. It could result in significant voltage differences between the powered device and local ground.
I have seen at most about 5 volts between neutral and ground. Leakage currents. I don't see any danger in the setup, at least from an "electrical" point of view. Electronics and communications... that's another thing. But Neutral is isolated from ground and chassis on all equipment. On the other hand, all houses in Spain have a residual current tripper, whatever the correct name is. If the current on live and neutral wires are slightly different (20 mA, I think), it triggers and switches of both wires. Ie, if live wire touches the chassis, or a person touches the wires, because some of the current goes back via earth instead of the other wire. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 07/08/17 23:02, John Andersen wrote:
On 08/07/2017 02:02 PM, Carlos E. R. wrote:
I'm not familiar with the British electricity code. I guess it will be different! Yes, start with that. (The not familiar part).
There are a ton of differences in EU and North American power systems as far as physical structure. Its easy to be less familiar than you think you are. This stuff has become more than a little complex with the growth of grids.
Neutral is tied to Ground almost everywhere in the US, if not on your house then on the next nearest pole or power pedestal. For me it is via the sub-station at the end of the street. Then there are differences between three wire systems (pos, neg, neut) and 4 wire systems (pos net neut ground) that occur in different parts of the world. Some parts of the US have individual houses all on one side (phase), and alternate every other house to the other leg to balance the phases. Then someone strings an extension cord. (Indications are that the US may slowly be moving toward a 4 wire infrastructure. https://en.wikipedia.org/wiki/Three-phase_electric_power We are on 3 phase electric in the UK. But as far as These Tape Drives, this seems unlikely to be a ground issue unless some parts of the tape backup system are on the opposite leg. Electrical code frowns on this and wants all outlets in the same room to be on the same leg.
There could also be large ground differences if some parts of the tape system are on a UPS and the other parts are Not, or are on a different UPS. Everything on this system is connected via a single AC BR1500 GI UPS on one plug socket.
A multi-meter can usually detect this situation for by checking for voltage between ground sockets between outlets. I checked this and there seems to be no difference. I also checked all
To me it seems unlikely to be a ground issue as my HP 448 LTO2 / 232 LTO1 / HP DAT72 Sony AIT drives all work perfectly but I cannot get either of the HP 1760 LTO4 drives to work?? the screw connections in the sockets and in the consumer unit and they are all good. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Paul, I'd chase the SCSI error you're now getting. It is consistent/reliable, unlike the problems I was fighting 25 years ago. FYI only: When I had my power issues in France 25 years ago I found nothing wrong when testing with a voltmeter. But I was getting desperate and rented a AC line monitor that was the size of an oscilloscope or bigger. Its job was to look for spurious anomalies and record them. As I recall, it detected significant spikes on the ground/neutral throughout the day as other equipment in the building turned on and off. It was a commercial building with 50+ people working in it, but of a low-end modular construction. As I recall, each room had a window A/C unit. I don't know EU power well enough to knowledgeably discuss how spikes could happen or how it could impact the computer's operation. Maybe there were other ways to fix the problem. I only know we fixed it by installing the ground rod. This picture is of a similar cockpit trainer, but I don't recall what kind of airplane the cockpit trainer was for: http://www.flightdecksolutions.com/assets/images/large_images/fds-a320-pro-m... I installed cockpit trainers for numerous commercial airplanes (Boeing/Mcdonnell Douglas/Airbus) during the decade I worked in that space. All the ones I installed were for airplanes with "glass cockpits" (see the picture at the link.) and we simulated the glass instruments, so we had numerous computers in the solution. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 14:41, Greg Freemyer wrote:
Paul,
I'd chase the SCSI error you're now getting. It is consistent/reliable, unlike the problems I was fighting 25 years ago.
FYI only:
When I had my power issues in France 25 years ago I found nothing wrong when testing with a voltmeter. But I was getting desperate and rented a AC line monitor that was the size of an oscilloscope or bigger. Its job was to look for spurious anomalies and record them. As I recall, it detected significant spikes on the ground/neutral throughout the day as other equipment in the building turned on and off. It was a commercial building with 50+ people working in it, but of a low-end modular construction. As I recall, each room had a window A/C unit. I don't know EU power well enough to knowledgeably discuss how spikes could happen or how it could impact the computer's operation. Maybe there were other ways to fix the problem. I only know we fixed it by installing the ground rod.
Any motor starting-stopping with switches (like a thermostat) produces a spike on the lines, but not on the ground line. However, line filters pass on the spikes to the ground line. If the building ground line (I'm considering the UK or France electricity code similar in this respect to Spain code) is not good, ie, there is a noticeable resistance, then the voltage in the ground line goes up in all the building. So in that case using a separate ground does a difference.
This picture is of a similar cockpit trainer, but I don't recall what kind of airplane the cockpit trainer was for: http://www.flightdecksolutions.com/assets/images/large_images/fds-a320-pro-m...
I want one for home! ;-) -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/07/2017 04:02 PM, Carlos E. R. wrote:
So unless something is causing some other type of interference on the earth? Out earth connection is via the Neutral cable coming into the building.
I think Paul means absent being located over (or in near proximity to) the Large Hadron Collider ...
I don't understand this sentence. You don't have a separate ground wire, instead you use the neutral cable?
The building doesn't have it's own ground pole, a metal rod inserted into the earth?
It depends. Older buildings may, I know at my office I ended up driving an 8-foot by 1/2-inch copper rod into the ground beneath the server closet to provide a direct ground. (thankfully there was an access to the crawl-space in the server closet itself....)
I'm not familiar with the British electricity code. I guess it will be different! :-p
I've always thought those UK adapters you got with equipment meant for international use -- looked a little dodgy... -- David C. Rankin, J.D.,P.E.
On 07/08/17 22:02, Carlos E. R. wrote:
On 2017-08-07 19:43, Paul Groves wrote:
So unless something is causing some other type of interference on the earth? Out earth connection is via the Neutral cable coming into the building. I don't understand this sentence. You don't have a separate ground wire, instead you use the neutral cable?
The building doesn't have it's own ground pole, a metal rod inserted into the earth?
I'm not familiar with the British electricity code. I guess it will be different! :-p I would post a scan of the page detailing this but it will not let me attach an image here.
You can have one of 3 types of earth in the UK as far as I understand. We have the first type where the metal shielding around the live (or phase) is used as ground/neutral. Or there is sometimes a separate shared neutral/ground cable. Or if none of the above a metal ground spike in the ground under the building (this is rare in the UK) for houses).
We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.). They need to measure it themselves with their own equipment, and I suppose they are very hard to convince to do the measurement. We have asked but they ignored us. I have got 6 months of the voltage logged in an sql database. I am planning to generate a graph from this so we can claim back the costs. Perhaps this could be causing the problem? Dunno.
I have yet to transport the system to a different site to test it. If that worked, it would nail the problem, certainly.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/08/17 22:02, Carlos E. R. wrote:
On 2017-08-07 19:43, Paul Groves wrote:
So unless something is causing some other type of interference on the earth? Out earth connection is via the Neutral cable coming into the building.
I don't understand this sentence. You don't have a separate ground wire, instead you use the neutral cable?
The building doesn't have it's own ground pole, a metal rod inserted into the earth?
I'm not familiar with the British electricity code. I guess it will be different! :-p
Yup! For a domestic building we have single-phase 230V supply - NOT 240V as many brits believe! The European standard is 230V, and we moved to the standard maybe 20 years ago? (The standard says that the deviation permitted goes up to the old UK 240V, and down to other countries' old 210V or 220V, so all appliances must be able to cope with everything, but the supply is supposed to be 230V.) Neutral is bonded to ground AT THE SUBSTATION (to the best of my knowledge). All of our electrics are 3-wire, live, neutral and ground, with ground usually bonded - in the house - to the domestic water system. Sounds dangerous until you realise pipes are mostly copper so it's actually a massive electrical sink, and pretty safe. Dunno how it changes things now we're moving to plastic pipes, but I'd have thought it doesn't actually make much difference. Water is a conductor. The other thing is, ALL modern domestic wiring runs through circuit breakers, with an earth leakage trip. And these things are damn sensitive. 30mA leaking to earth, and the inbound supply trips, taking out the entire house electrical supply. So if you cut through your electric lawn mower's cable, you won't get a shock. Everybody in the house will, though, as the trip takes out all power! That's why, in a previous house, I had super-sensitive socket trips fitted in the conservatory - 5mA - so that a garden accident wouldn't take the house out. And all sockets are L/N/E with safety blocks - there's a safety plate that prevents access to live or neutral, until the plug's long earth pin pushes it out of the way. If your appliance doesn't have an earth, the cable still needs a non-functional earth pin in order to allow you to plug it in. I remember a guy on Groklaw - who knew a bit about electrics - being horrified by the french system where they have both L/N/E and L/N sockets - and apparently you can put an L/N/E plug into an L/N socket (sounds dangerous!), but not an L/N plug into an L/N/E socket (where's the harm in that?). Apparently it's perfectly safe, for other reasons, but that's well weird to us brits. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Dunno how it changes things now we're moving to plastic pipes, but I'd have thought it doesn't actually make much difference. Water is a conductor.
The US has local building codes, but where I live if you use the cold water line for ground/earth, you have to do it close to where the line comes in the building (5 ft? 1.6 meters?) If you do it randomly throughout a house, you may indeed end up with plastic pipe in the way, and that is frowned upon. (Against the building code.) Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/11/2017 04:25 PM, Greg Freemyer wrote:
The US has local building codes, but where I live if you use the cold water line for ground/earth, you have to do it close to where the line comes in the building (5 ft? 1.6 meters?)
If you do it randomly throughout a house, you may indeed end up with plastic pipe in the way, and that is frowned upon. (Against the building code.)
Not to mention create several distinct 'ground loop' within the system.... -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Fri, 11 Aug 2017 22:15:45 +0100 Wols Lists <antlists@youngman.org.uk> wrote:
On 07/08/17 22:02, Carlos E. R. wrote:
On 2017-08-07 19:43, Paul Groves wrote:
So unless something is causing some other type of interference on the earth? Out earth connection is via the Neutral cable coming into the building.
I don't understand this sentence. You don't have a separate ground wire, instead you use the neutral cable?
The building doesn't have it's own ground pole, a metal rod inserted into the earth?
I'm not familiar with the British electricity code. I guess it will be different! :-p
Yup!
For a domestic building we have single-phase 230V supply - NOT 240V as many brits believe! The European standard is 230V, and we moved to the standard maybe 20 years ago? (The standard says that the deviation permitted goes up to the old UK 240V, and down to other countries' old 210V or 220V, so all appliances must be able to cope with everything, but the supply is supposed to be 230V.)
Well, yes and no. The standard is harmonised as you say and equipment manufactured now must be able to deal with the limits. But the voltage supplied depends on the distribution transformers and those are replaced as it becomes necessary, not just because there was a European harmonisation. That's why the tolerance band in the standard is so large. So in some places in the UK the nominal supply voltage is still 240 V just as in some places on the continent it is still 220 V.
Neutral is bonded to ground AT THE SUBSTATION (to the best of my knowledge). All of our electrics are 3-wire, live, neutral and ground, with ground usually bonded - in the house - to the domestic water system. Sounds dangerous until you realise pipes are mostly copper so it's actually a massive electrical sink, and pretty safe. Dunno how it changes things now we're moving to plastic pipes, but I'd have thought it doesn't actually make much difference. Water is a conductor.
Again, yes and no. Most houses are supplied by an underground supply that includes live, neutral and CPC (earth), grounded as you say at the source. But some houses, mostly a lot of those supplied by overhead wires, only have two incoming wires - live and neutral. The neutral is earthed at the supply and in some cases is fed to the house as both neutral and earth. In other cases it is fed only as neutral and the consumer must provide their own earth using a ground spike. Bonding to the water pipe is still very common but is not strictly necessary in the now very common case that the incoming water main is plastic. Bonding of metal taps etc is also no longer required. For more detail see: http://electrical.theiet.org/wiring-matters/16/earthing-questions.cfm?type=p...
The other thing is, ALL modern domestic wiring runs through circuit breakers, with an earth leakage trip. And these things are damn sensitive. 30mA leaking to earth, and the inbound supply trips, taking out the entire house electrical supply. So if you cut through your electric lawn mower's cable, you won't get a shock. Everybody in the house will, though, as the trip takes out all power!
It's not true that all circuits run through RCDs (residual current devices), which replaced earth leakage circuit breakers. Some circuits have to be protected and others can be. Alarm circuits are an example of a circuit that some people do not believe should be connected through an RCD (in case of a faulty RCD interfering with the alarm).
That's why, in a previous house, I had super-sensitive socket trips fitted in the conservatory - 5mA - so that a garden accident wouldn't take the house out.
And all sockets are L/N/E with safety blocks - there's a safety plate that prevents access to live or neutral, until the plug's long earth pin pushes it out of the way. If your appliance doesn't have an earth, the cable still needs a non-functional earth pin in order to allow you to plug it in.
I remember a guy on Groklaw - who knew a bit about electrics - being horrified by the french system where they have both L/N/E and L/N sockets - and apparently you can put an L/N/E plug into an L/N socket (sounds dangerous!), but not an L/N plug into an L/N/E socket (where's the harm in that?). Apparently it's perfectly safe, for other reasons, but that's well weird to us brits.
I don't think that's true but I'll let a French person respond. Cheers, Dave
Cheers, Wol
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-12 14:56, Dave Howorth wrote:
Again, yes and no. Most houses are supplied by an underground supply that includes live, neutral and CPC (earth), grounded as you say at the source. But some houses, mostly a lot of those supplied by overhead wires, only have two incoming wires - live and neutral. The neutral is earthed at the supply and in some cases is fed to the house as both neutral and earth. In other cases it is fed only as neutral and the consumer must provide their own earth using a ground spike.
I have my doubts about that. In Spain (I doubt that Britain is that different in this respect) there are only two wires entering the house, live and neutral, no matter if it is underground pipes or aerial. The Earth wire is tied to a metal rod (or several) "nailed" on to the earth. Using the water system is no longer permitted (because not all the pipes are metal, because it can cause corrosion, because voltages can develop...) Tying the earth to the neutral I consider dangerous. There is no redundancy. But we should keep all replies on the offtopic mail list only. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
So I have discovered that the problem lies somewhere between my PSU and the LTO4 tape drives. The PSU An Antec 80 Plus gold rated 1000W supply. At peak power draw (on boot) the server draws 650 -ish Watts, during operation no more than 250W, so the power supply is not under heavy load. All of the voltages on each rail measure within 0.15V of the rated voltage. So when I start to write to the LTO4 drive, it appears it is re-setting itself due to a power problem? this happens on both drives as I said before. Recap: The drive writes for a few seconds. Goes 'donk' then the orange LEDs light up and the drive reboots. At this point I have to reboot the server (probably because SCSI is not hot-swappable). So I connected an external PSU only to the LTO4 drives (one at a time) while they are still connected to the server via SCSI and they seems to work fine. So I have spent hours swapping leads / cleaning connections / measuring resistance on my PSU and all its cables but I cannot see anything obviously wrong. I even disconnected every item of hardware apart from the LTO4 drive and an ubuntu live USB. The problem still occurs. So I temporarily installed a lower power PSU into the server (because all of the hard drives and other tape drives are connected). And repeated the test with the ubuntu live usb. And the drives work fine. so the remaining puzzle is: why won't the LTO 4 drives work with my expensive Antec Server PSU but they do work with a £7.50 cheapo unbranded chinese supply? And also why do all my other tape drives work fine with the Antec PSU? Is it a problem with the model (HP 1760 SCSI)? But they work fine on another PSU so the LTO4 drives cannot be at fault can they? Anyone have a clue how to proceed? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 30/08/17 11:54 AM, Paul Groves wrote:
Anyone have a clue how to proceed?
Hypothesising .... Use an oscilloscope, preferably one with storage/playback instead of a meter to measure the voltage in each case. it may be something to do with waveforms, leakage of higher frequencies from the switch-mode power supply. It may be that the smoothing capacitors that are supposed to filter out those frequencies are doing other thing when there is a surge, releasing initial power, so the drive starts, but not following it up. A storage scope will let you see this. BTDT but for different equipment, different context (avionics). Differences in design and implementation. Some other manufacturer's el-cheapo PSU might be done differently from the el-cheapo one that works. Maybe you got the luck of the draw. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-30 17:54, Paul Groves wrote:
So I have discovered that the problem lies somewhere between my PSU and the LTO4 tape drives. The PSU An Antec 80 Plus gold rated 1000W supply.
...
Recap: The drive writes for a few seconds. Goes 'donk' then the orange LEDs light up and the drive reboots. At this point I have to reboot the server (probably because SCSI is not hot-swappable).
...
So I temporarily installed a lower power PSU into the server (because all of the hard drives and other tape drives are connected). And repeated the test with the ubuntu live usb. And the drives work fine.
Wow :-O
so the remaining puzzle is: why won't the LTO 4 drives work with my expensive Antec Server PSU but they do work with a £7.50 cheapo unbranded chinese supply? And also why do all my other tape drives work fine with the Antec PSU? Is it a problem with the model (HP 1760 SCSI)? But they work fine on another PSU so the LTO4 drives cannot be at fault can they?
Anyone have a clue how to proceed?
Anton hypothesis is the best bet, I think. Using a scope is a good idea. If that turns out to be true, it may be that the older tapes have better internal power filtering while the new ones rely on the PSU. Or it may be that the socket on the new ones doesn't fit well with the old psu and do better with the cheap psu. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/30/2017 08:54 AM, Paul Groves wrote:
The PSU An Antec 80 Plus gold rated 1000W supply.
At peak power draw (on boot) the server draws 650 -ish Watts, during operation no more than 250W, so the power supply is not under heavy load. All of the voltages on each rail measure within 0.15V of the rated voltage.
Why not contact Antec? This may be a known problem, or the result of degrading capacitors inside. The good (sort of) news is they also make a higher powered power unit. http://www.antec.com/product.php?id=706569&fid=343 of 1300 watts continuous. However, I'd think twice of throwing more money at them if their products have a problem supporting less than half their rated draw. Amazon wants 300 bucks for that model, but maybe antec can be bullied into offering it to you in exchange for your current one rather than suffer bad press. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Wed, Aug 30, 2017 at 11:54 AM, Paul Groves <paul.groves.787@gmail.com> wrote:
so the remaining puzzle is: why won't the LTO 4 drives work with my expensive Antec Server PSU but they do work with a £7.50 cheapo unbranded chinese supply? And also why do all my other tape drives work fine with the Antec PSU? Is it a problem with the model (HP 1760 SCSI)? But they work fine on another PSU so the LTO4 drives cannot be at fault can they?
Anyone have a clue how to proceed?
Paul, It was me suggesting ground/power issues, so I'm not shocked at your finding. Are you using SE signalling or differential on your SCSI cable? On setups with numerous SCSI tape drives I used differential. If you do that you can power the drives from a totally different PSU. I personally have never put more than 1 SE tape drive on a SCSI cable. Purely by chance, whenever I had numerous SCSI tape drives hooked up, I put the drives in external enclosures with their own PSU and used differential signaling. That doesn't tell you why your drives are incompatible with your PSU, but if you're simply tired of messing with it, it may give you a way forward. Good Luck, Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/31/2017 12:25 PM, Greg Freemyer wrote:
I personally have never put more than 1 SE tape drive on a SCSI cable. Purely by chance, whenever I had numerous SCSI tape drives hooked up, I put the drives in external enclosures with their own PSU and used differential signaling.
The external enclosures we used ONLY supplied external power. As far as data cables, they were dumb as rocks, and everything was on the single SCSI cable. We had two tape drives and two disk drives in the external. I don't remember ever making any change to signaling. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Thu, Aug 31, 2017 at 4:23 PM, John Andersen <jsamyth@gmail.com> wrote:
On 08/31/2017 12:25 PM, Greg Freemyer wrote:
I personally have never put more than 1 SE tape drive on a SCSI cable. Purely by chance, whenever I had numerous SCSI tape drives hooked up, I put the drives in external enclosures with their own PSU and used differential signaling.
The external enclosures we used ONLY supplied external power. As far as data cables, they were dumb as rocks, and everything was on the single SCSI cable. We had two tape drives and two disk drives in the external.
I don't remember ever making any change to signaling.
It seems all the newer SCSI versions are differential: https://en.wikipedia.org/wiki/Parallel_SCSI#Comparison_table All the SE (single ended) specs are 40MB/sec or slower. The Ultra-2 version is differential and it was introduced in 1997. So, there's a good chance you were using differential signalling even if you didn't know it. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, 7 Aug 2017 18:43:06 +0100 Paul Groves <paul.groves.787@gmail.com> wrote:
We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.).
Do you or any other properties nearby and on the same phase have solar panels (PV)? They will cause the mains voltage to be higher than normal. The voltage limits are Europe-wide, incidentally. They were 'harmonised' a while ago. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 00:17, Dave Howorth wrote:
On Mon, 7 Aug 2017 18:43:06 +0100 Paul Groves <paul.groves.787@gmail.com> wrote:
We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.).
Do you or any other properties nearby and on the same phase have solar panels (PV)? They will cause the mains voltage to be higher than normal.
The voltage limits are Europe-wide, incidentally. They were 'harmonised' a while ago.
To 230, I think. The UK is still largely at 240. Germany... I seem to recall had a higher voltage. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/08/17 00:53, Carlos E. R. wrote:
On 2017-08-08 00:17, Dave Howorth wrote:
We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.). Do you or any other properties nearby and on the same phase have solar
On Mon, 7 Aug 2017 18:43:06 +0100 Paul Groves <paul.groves.787@gmail.com> wrote: panels (PV)? They will cause the mains voltage to be higher than normal.
The voltage limits are Europe-wide, incidentally. They were 'harmonised' a while ago. To 230, I think. The UK is still largely at 240. Germany... I seem to recall had a higher voltage.
We are still at 240V but the law in the UK is 230V +10% maximum (253V) -5% mimimum (218.5V) I personally think 240 makes more sense as it would always require an even number of turns on the transformer. Also, especially in pre-calculator days as it is divisible by 12 so easy to calculate. (e.g. 240v to 12v is a 20:1 transformer). Simple eh.. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/08/2017 08:12 AM, Paul Groves wrote:
We are still at 240V but the law in the UK is 230V +10% maximum (253V) -5% mimimum (218.5V)
I personally think 240 makes more sense as it would always require an even number of turns on the transformer. Also, especially in pre-calculator days as it is divisible by 12 so easy to calculate. (e.g. 240v to 12v is a 20:1 transformer). Simple eh..
That ignores a real factor in power distribution, line loss. As you get further from the source,the line voltage will decrease. This can be compensated for, to a degree, by using taps on the transformer. But even then, changing loads will cause changing voltage. This is why the voltage drops at peak demand times. So, everything is designed for nominal voltages, with the expectation there will be variations. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/08/17 08:41 AM, James Knott wrote:
That ignores a real factor in power distribution, line loss. As you get further from the source,the line voltage will decrease. This can be compensated for, to a degree, by using taps on the transformer. But even then, changing loads will cause changing voltage. This is why the voltage drops at peak demand times. So, everything is designed for nominal voltages, with the expectation there will be variations.
From the consumers POV it means that more power is being consumed to get the same amount of work done. When the voltage and current are completely out of
That depends on a LOT of things. Power _distribution_ is usually at a higher voltage since the Joule–Lenz law means that resistive line losses are based on current. A local 'substation' has the transformer that drops the voltage to that required for the building. Domestic and industrial may not be the same. It's more efficient to push that final stage as close to the 'load' as possible. But there's another factor that comes into play. We're talking about A/C -- that's so we can play with the voltage via transformers, thank you very much Tesla and Westinghouse. But few A/C loads are purely resistive. Motors, which dominate, present and inductive load. There are also devices that present a capacitive load, but motors predominate. This generates a phase shift so that the voltage and current vectors are no longer at 90 degrees. This causes a power loss. It is also pushed back towards the generator and can damage plant and equipment. phase no useful work is being done even though the power is being transmitted. There are ways to ameliorate this. One is to put a honking great capacitor across the motor. You see this done with the so-called 'energy efficient' domestic refrigerators at Sears. There are also firms that sell capacitor banks to retrofit older equipment such as refrigerated display cabinets in stores and the walk-in fridges in some restaurants and warehouses. Yes, there are refrigerator mechanism that don't use electric motors; Einstein invented one. But they aren't very popular. If you look at your (even domestic) electricity bill it probably has a footnote about the assumed phase shift. Some industrial plants may have a phase shift monitor. Some of those may be tied to an alarm. The use of three-phase makes phase shift easier to handle under many conditions. Three phase motors offer more possibilities for load balancing. Some domestic settings might have a 3-phase feed for a street with alternating taps supplying what amount to single phase for individual houses. Back in my university-vacation days I worked as an 'apprentice' to a firm of electricians upgrading older houses in England, houses built between the wars or just after the 'last war'. I learnt a lot about the 'Code' and what was the minimum standard and what was a sensible way of re-wiring and bring those houses up to the modern standard. Honestly, even by the UK 'Code' of the 1970s I find the electrical code here in Canada falls short. IIR the later upgrade to the German-driven EU standard for electrical installation and wiring components was even more vigorous. But back then I was shown a fair deal about the domestic/household end of the power distribution scheme worked. The gap between 2-phase and 3-phase was an important aspect of that. The threshold of what a single house could pull of a phase before it had to be converted to 3-phase was another. Wiring a house that had 3-phase coming in but had room-level 2-phase because the appliances were all 2-phase added complications. In a industrial/factory setting that wasn't an issue for the same reason its not an issue on a normal street or in a condominium hi-rise. The 'peeling off' of individual phases, local transformers & taps, is more manageable in those settings. In a single house, an old Victorian monolith that had been converted to individual flats where each one had to have its own meters or a between-the-wars cottage that had been extended and now had a a hi-tech 'entertainment centre' that was sucking power, figuring out a 'balance' is another matter. But for many settings, phases losses of power WITHOUT any voltage drop are the issue. https://en.wikipedia.org/wiki/AC_power#Active.2C_reactive.2C_and_apparent_po... https://en.wikipedia.org/wiki/Power_factor https://en.wikipedia.org/wiki/Power_factor#Importance_of_power_factor_in_dis... And of relevance to us: <quote src=:https://en.wikipedia.org/wiki/Power_factor#Switched-mode_power_supplies"> A particularly important class of non-linear loads is the millions of personal computers that typically incorporate switched-mode power supplies (SMPS) with rated output power ranging from a few watts to more than 1 kW. Historically, these very-low-cost power supplies incorporated a simple full-wave rectifier that conducted only when the mains instantaneous voltage exceeded the voltage on the input capacitors. This leads to very high ratios of peak-to-average input current, which also lead to a low distortion power factor and potentially serious phase and neutral loading concerns. A typical switched-mode power supply first converts the AC mains to a DC bus by means of a bridge rectifier or a similar circuit. The output voltage is then derived from this DC bus. The problem with this is that the rectifier is a non-linear device, so the input current is highly non-linear. That means that the input current has energy at harmonics of the frequency of the voltage. This presents a particular problem for the power companies, because they cannot compensate for the harmonic current by adding simple capacitors or inductors, as they could for the reactive power drawn by a linear load. Many jurisdictions are beginning to legally require power factor correction for all power supplies above a certain power level. Regulatory agencies such as the EU have set harmonic limits as a method of improving power factor. Declining component cost has hastened implementation of two different methods. To comply with current EU standard EN61000-3-2, all switched-mode power supplies with output power more than 75 W must include passive power factor correction, at least. 80 Plus power supply certification requires a power factor of 0.9 or more.[ </quote> -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 16:33, Anton Aylward wrote:
This generates a phase shift so that the voltage and current vectors are no longer at 90 degrees. This causes a power loss. It is also pushed back towards the generator and can damage plant and equipment.
From the consumers POV it means that more power is being consumed to get the same amount of work done. When the voltage and current are completely out of phase no useful work is being done even though the power is being transmitted.
More current, no more power. When the phase shift is 90° the power is zero.
there are refrigerator mechanism that don't use electric motors; Einstein invented one. But they aren't very popular.
Because they are very inefficient and expensive. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/08/2017 10:42 AM, Carlos E. R. wrote:
From the consumers POV it means that more power is being consumed to get the
same amount of work done. When the voltage and current are completely out of phase no useful work is being done even though the power is being transmitted. More current, no more power. When the phase shift is 90° the power is zero.
Of course, there aren't a lot of loads that are purely reactive. Also, for many years, fluorescent tubes used inductors for ballast, which introduced significant power factor. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/08/17 10:50 AM, James Knott wrote:
On 08/08/2017 10:42 AM, Carlos E. R. wrote:
From the consumers POV it means that more power is being consumed to get the
same amount of work done. When the voltage and current are completely out of phase no useful work is being done even though the power is being transmitted. More current, no more power. When the phase shift is 90° the power is zero.
Of course, there aren't a lot of loads that are purely reactive.
Of course not. Everything has wiring so there is always _some_ resistive loss. But a motor WORKS by induction!
Also, for many years, fluorescent tubes used inductors for ballast, which introduced significant power factor.
That's 'balancing' factor. The tubes by themselves represent a capacitive load at low frequencies. At the higher kilohertz frequencies that Tesla operated the, that's another matter. There's a long and ugly history of why we have such a low frequency for power distribution. Be glad its not 15Hz or lower! -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/08/2017 11:39 AM, Anton Aylward wrote:
There's a long and ugly history of why we have such a low frequency for power distribution. Be glad its not 15Hz or lower!
I'm familiar with some of it. How higher frequencies were better for lighting, but lower frequencies were easier to generate at the time, power line vs iron losses, etc. Incidentally, at one time 25 Hz power was common in North America. That came about because of the first generating plant at Niagara. When they decided to go with AC, a lot of the mechanical details, such as turbines had already been fixed and they were pretty much limited to 25 Hz AC. It wasn't that long ago, that the last of the 25 Hz generation at Niagara was shut down. However, there were a lot of different frequencies used years ago. Also, in Newfoundland, there were some parts of the province on 50 Hz, with the rest on 60. I believe that situation still exists in Japan. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Seems to me that this thread has gone sufficiently off the rails as to be of no help to the original question, and it's almost time for Mr Brown to weigh in with threats. -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/08/17 17:54, John Andersen wrote:
Seems to me that this thread has gone sufficiently off the rails as to be of no help to the original question, and it's almost time for Mr Brown to weigh in with threats.
Agreed (although i have no idea who Mr. Brown is), so to get back on track: I ran mt -f /dev/st0 erase 0 the drive started to erase, at the point it started writing to the tape the same problem occurs. The same happens using tar or HP LTT to write to the tape. The drive goes 'donk' and stops dead. The tape spools for some time back and forth and repeats occasionally 'donk'ing. Eventually the drive gives up and ejects the tape then takes it back in. This repeats forever until I eject the cartridge. The server at this point is frozen on the command still. I cannot Ctrl + C, Ctrl + Z or Ctrl + \ I have no choice but to force reboot using sysrq keys. The same has happened with both of my LTO4 drives. Both HP Storageworks 1760 If I go to HP LTT, select the drive and look at the health report, It has logged the following two messages every time I experience the problem: ISCSI TaskMgmtFunction invoked: SI_LOGICAL_UNIT_RESET_MGMT_FUNC, Initiator ID: 0000h, Via SI_PRIMARY_PORT_ID, To LUN: 0h BusRst ISCSI Bus Reset Occured I have no clue what to do at this point. Anyone have any ideas? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 19:44, Paul Groves wrote:
If I go to HP LTT, select the drive and look at the health report, It has logged the following two messages every time I experience the problem:
ISCSI TaskMgmtFunction invoked: SI_LOGICAL_UNIT_RESET_MGMT_FUNC, Initiator ID: 0000h, Via SI_PRIMARY_PORT_ID, To LUN: 0h
BusRst ISCSI Bus Reset Occured
I have no clue what to do at this point. Anyone have any ideas?
Is there any way to ask HP for support? -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/08/17 20:37, Carlos E. R. wrote:
On 2017-08-08 19:44, Paul Groves wrote:
If I go to HP LTT, select the drive and look at the health report, It has logged the following two messages every time I experience the problem:
ISCSI TaskMgmtFunction invoked: SI_LOGICAL_UNIT_RESET_MGMT_FUNC, Initiator ID: 0000h, Via SI_PRIMARY_PORT_ID, To LUN: 0h
BusRst ISCSI Bus Reset Occured
I have no clue what to do at this point. Anyone have any ideas? Is there any way to ask HP for support?
Not without spending lots of money on customer support. In my previous experience HP have always said they cannot do anything with a problem that involves anything technical and tell you to get a refund unless you spend MORE money on the extended warranty etc. So about as useful as a pedal-powered wheelchair. But I have had 3 different drives now all of the same model with the same problem. So I find it quite unlikely they are all experiencing the same fault. Especially as the previous owners all had them working fine. Perhaps firmware related? Although I am unable to write a firmware tape to update them :( I could try on a windows server to rule out the linux driver perhaps. Yuk! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-09 01:10, Paul Groves wrote:
On 08/08/17 20:37, Carlos E. R. wrote:
I have no clue what to do at this point. Anyone have any ideas? Is there any way to ask HP for support?
Not without spending lots of money on customer support. In my previous experience HP have always said they cannot do anything with a problem that involves anything technical and tell you to get a refund unless you spend MORE money on the extended warranty etc. So about as useful as a pedal-powered wheelchair.
:-(
But I have had 3 different drives now all of the same model with the same problem. So I find it quite unlikely they are all experiencing the same fault. Especially as the previous owners all had them working fine.
All the same model, all fail? Well, then it is that model which is faulty.
Perhaps firmware related? Although I am unable to write a firmware tape to update them :(
I could try on a windows server to rule out the linux driver perhaps. Yuk!
I was going to suggest that. By all means, try a Windows machine. Even better, do it in the current machine, but running Windows, to rule out any hardware change from the equation. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 09/08/17 03:54, Carlos E. R. wrote:
On 2017-08-09 01:10, Paul Groves wrote:
On 08/08/17 20:37, Carlos E. R. wrote:
I have no clue what to do at this point. Anyone have any ideas? Is there any way to ask HP for support?
Not without spending lots of money on customer support. In my previous experience HP have always said they cannot do anything with a problem that involves anything technical and tell you to get a refund unless you spend MORE money on the extended warranty etc. So about as useful as a pedal-powered wheelchair. :-(
But I have had 3 different drives now all of the same model with the same problem. So I find it quite unlikely they are all experiencing the same fault. Especially as the previous owners all had them working fine. All the same model, all fail? Well, then it is that model which is faulty.
Perhaps firmware related? Although I am unable to write a firmware tape to update them :(
I could try on a windows server to rule out the linux driver perhaps. Yuk! I was going to suggest that. By all means, try a Windows machine. Even better, do it in the current machine, but running Windows, to rule out any hardware change from the equation.
Well on windows the drive assesment test in LTT passed without errors. I think the next logical step would be to update the drive firmware but I am unable to download it because I do not have a HP Enterprise account. Could someone with an account please help me out by downloading the firmware for my drives and emailing it to me (only 1.4MB apparently)? I need version W62D for EH921A (Storageworks Ultrium 1760 SCSI LTO4) h20564.www2.hpe.com/hpsc/swd/public/detail?swItemId=co_99740_1 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Does anyone know of a free bit of software to write some files to tape for Windows (Like tar)? Or does someone have a spare Symantec Backup Exec license I could borrow to test my drives? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Wed, Aug 9, 2017 at 1:51 PM, Paul Groves <paul.groves.787@gmail.com> wrote:
Does anyone know of a free bit of software to write some files to tape for Windows (Like tar)?
Or does someone have a spare Symantec Backup Exec license I could borrow to test my drives?
Paul, Have you tried the included "Backup and Restore" tool: start->control panel->backup and restore I'll be disappointed if they've removed tape support from it. Back in Windows 2000/2003 days it was caused NTbackup. I think the above is just a rename. I know in 2003 it could write to backup tapes. Alternatively, you could try to install cygwin and use its version of tar. I use cygwin all the time, but I haven't tried it with a tape drive in many years. Good Luck Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 08/09/2017 01:51 PM, Paul Groves wrote:
Does anyone know of a free bit of software to write some files to tape for Windows (Like tar)?
Or does someone have a spare Symantec Backup Exec license I could borrow to test my drives?
There is tar in Cygwin and also the new Linux in Windows 10. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-10 02:45, James Knott wrote:
On 08/09/2017 01:51 PM, Paul Groves wrote:
Does anyone know of a free bit of software to write some files to tape for Windows (Like tar)?
Or does someone have a spare Symantec Backup Exec license I could borrow to test my drives?
There is tar in Cygwin and also the new Linux in Windows 10.
But the idea is to find out if the devices work properly in native Windows. If they fail with Cygwin you prove nothing. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 09/08/17 19:43, Greg Freemyer wrote:
On Wed, Aug 9, 2017 at 1:51 PM, Paul Groves <paul.groves.787@gmail.com> wrote:
Does anyone know of a free bit of software to write some files to tape for Windows (Like tar)?
Or does someone have a spare Symantec Backup Exec license I could borrow to test my drives? Paul,
Have you tried the included "Backup and Restore" tool:
start->control panel->backup and restore They disabled that in 2008r2 (but you could enable it with a load of registry teaks. They then removed it in 2012 onward saying "support for legacy hardware is no longer included".
One of many reasons I have no windows servers in production use. On 10/08/17 01:48, Carlos E. R. wrote:
On 2017-08-10 02:45, James Knott wrote:
On 08/09/2017 01:51 PM, Paul Groves wrote:
Does anyone know of a free bit of software to write some files to tape for Windows (Like tar)?
Or does someone have a spare Symantec Backup Exec license I could borrow to test my drives?
There is tar in Cygwin and also the new Linux in Windows 10. But the idea is to find out if the devices work properly in native Windows. If they fail with Cygwin you prove nothing.
That is a good point. I guess it would prove that the windows driver is working but it might fail in the emulated linux environment. I will try this anyway to see what happens. In the meantime is there a free bit of windows software to write to tape? Paul -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-10 10:28, Paul Groves wrote:
On 09/08/17 19:43, Greg Freemyer wrote:
On Wed, Aug 9, 2017 at 1:51 PM, Paul Groves <> wrote:
There is tar in Cygwin and also the new Linux in Windows 10. But the idea is to find out if the devices work properly in native Windows. If they fail with Cygwin you prove nothing.
That is a good point. I guess it would prove that the windows driver is working but it might fail in the emulated linux environment. I will try this anyway to see what happens.
In the meantime is there a free bit of windows software to write to tape?
Last tape I used was around 1997, and I used some proprietary software included with the unit. I understand you got your units second hand; it is possible that the initial owner kept the software, if it existed. Otherwise, maybe the manufacturer provides some software for free :-? It might be not the full fledged one. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/10/2017 03:28 AM, Paul Groves wrote:
Paul,
Have you tried the included "Backup and Restore" tool:
start->control panel->backup and restore They disabled that in 2008r2 (but you could enable it with a load of registry teaks. They then removed it in 2012 onward saying "support for legacy hardware is no longer included".
That looks like: https://technet.microsoft.com/en-us/library/cc753528%28v=ws.11%29.aspx?f=255&MSPPError=-2147217396 Free/virusware (may be fine -- scan scan scan): http://www.z-dbackup.com/ https://www.iperiusbackup.com/ 49 Tools?: http://www.enterprisestorageforum.com/backup-recovery/open-source-storage-49... Here is one with a free trial: https://www.uranium-backup.com/uranium-backup-free-download/ -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 16:50, James Knott wrote:
On 08/08/2017 10:42 AM, Carlos E. R. wrote:
From the consumers POV it means that more power is being consumed to get the
same amount of work done. When the voltage and current are completely out of phase no useful work is being done even though the power is being transmitted. More current, no more power. When the phase shift is 90° the power is zero.
Of course, there aren't a lot of loads that are purely reactive. Also, for many years, fluorescent tubes used inductors for ballast, which introduced significant power factor.
In Spain, all traditional fluorescent tubes use inductors for ballast, since decades. The "new" compacts don't, they use electronic switching. A theoretical coil is purely reactive, but such a thing doesn't exist, unless you can manage to use superconductors :-p -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/08/17 10:42 AM, Carlos E. R. wrote:
On 2017-08-08 16:33, Anton Aylward wrote:
This generates a phase shift so that the voltage and current vectors are no longer at 90 degrees. This causes a power loss. It is also pushed back towards the generator and can damage plant and equipment.
From the consumers POV it means that more power is being consumed to get the same amount of work done. When the voltage and current are completely out of phase no useful work is being done even though the power is being transmitted.
More current, no more power. When the phase shift is 90° the power is zero.
Right. And it really buqqers up the generator and other equipment!
there are refrigerator mechanism that don't use electric motors; Einstein invented one. But they aren't very popular.
Because they are very inefficient and expensive.
There are many who contest that. At the time, material technology and manufacturing technique was limiting. Revising that today can make it much more reliable and much more efficient. More to the point, it doesn't need electricity. Any heat source will do. https://www.theguardian.com/science/2008/sep/21/scienceofclimatechange.clima... There's a point where the notional 'efficiency' really doesn't matter. https://www.scientificamerican.com/article/solar-refrigeration/ I bet those fishermen in the village of Maruata don't worry about 'efficiency' of their ice makers. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2017-08-08 17:34, Anton Aylward wrote:
On 08/08/17 10:42 AM, Carlos E. R. wrote:
there are refrigerator mechanism that don't use electric motors; Einstein invented one. But they aren't very popular.
Because they are very inefficient and expensive.
Replying on the offtopic mail list. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 2017-08-08 14:12, Paul Groves wrote:
On 08/08/17 00:53, Carlos E. R. wrote:
On 2017-08-08 00:17, Dave Howorth wrote:
On Mon, 7 Aug 2017 18:43:06 +0100 Paul Groves <> wrote:
The voltage limits are Europe-wide, incidentally. They were 'harmonised' a while ago. To 230, I think. The UK is still largely at 240. Germany... I seem to recall had a higher voltage.
We are still at 240V but the law in the UK is 230V +10% maximum (253V) -5% mimimum (218.5V)
I personally think 240 makes more sense as it would always require an even number of turns on the transformer. Also, especially in pre-calculator days as it is divisible by 12 so easy to calculate. (e.g. 240v to 12v is a 20:1 transformer). Simple eh..
230 is a compromise, so that all countries have the same voltage, thus facilitating travel and commerce. Spain had 220, going to 240 would need replacing most of the appliances. As it it, they changed the voltage silently, causing appliances, specially bulbs, to fail earlier. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 08/08/17 14:15, Carlos E. R. wrote:
On 2017-08-08 14:12, Paul Groves wrote:
On 08/08/17 00:53, Carlos E. R. wrote:
On 2017-08-08 00:17, Dave Howorth wrote:
On Mon, 7 Aug 2017 18:43:06 +0100 Paul Groves <> wrote:
The voltage limits are Europe-wide, incidentally. They were 'harmonised' a while ago. To 230, I think. The UK is still largely at 240. Germany... I seem to recall had a higher voltage.
We are still at 240V but the law in the UK is 230V +10% maximum (253V) -5% mimimum (218.5V)
I personally think 240 makes more sense as it would always require an even number of turns on the transformer. Also, especially in pre-calculator days as it is divisible by 12 so easy to calculate. (e.g. 240v to 12v is a 20:1 transformer). Simple eh..
230 is a compromise, so that all countries have the same voltage, thus facilitating travel and commerce. Spain had 220, going to 240 would need replacing most of the appliances. As it it, they changed the voltage silently, causing appliances, specially bulbs, to fail earlier.
As I understand it, they changed the standard a LOOONNGG time ago, and new appliances had to match the new European standard. So that was maybe 40 years ago. Then they changed the supply a fair while later, so by then the majority of old national standard stuff had already died and most kit actually extant could cope with the new voltage. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/08/17 23:17, Dave Howorth wrote:
We have had a problem lately where our voltage is always unusually high, hovering between 248V and 253V which is the absolute legal maximum they are allowed to supply in the UK. (They insist they are not providing this voltage but 3 UPSs and 4 volt-meteres would not all be false readings.). Do you or any other properties nearby and on the same phase have solar
On Mon, 7 Aug 2017 18:43:06 +0100 Paul Groves <paul.groves.787@gmail.com> wrote: panels (PV)? They will cause the mains voltage to be higher than normal. Never though of that. Although the UK voltage under law is 230V +- 5% for no more than 24 hours +10% maximum (so 253 maximum).
They are getting around this by setting the voltage back to 240.0 on a timer in the evenings. It is then back to 253V withing the hour. One time I saw it up to 268. This cannot be solar panels in the dark of course. They have over doubled the amount of properties in my village over the last few years but I have not seen any works to install another power feed. So logically if you want to draw more power you have to increase either the thickness of the cable (or number of cables) or the voltage. So personally I believe this is likely the cause of the high voltage. Although a lot of electronics in the UK are only made to stand 250V still (which is why I think everywhere keeps catching fire lately). Also the fact that there is no Bristish Standard for electrical goods any more and the CE electrical mark is self-regulated (which in my opinion in as good as no regulation).
The voltage limits are Europe-wide, incidentally. They were 'harmonised' a while ago.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 07/09/2017 11:16 AM, Paul Groves wrote:
On 09/07/17 17:23, Wols Lists wrote:
No that's not the question. A lot of SCSI devices come with built-in termination. Check the jumper documentation, and make sure all the internal terminators are disabled if you've got an external terminator.
If you have devices on all the blocks, there's nowhere to put a terminator which is why they do this, and which is why you should always have the terminator - or a terminating device - on the last block.
Cheers, Wol I can't see any mention in the manual of built in terminators. The jumper pins on all the drives are
Term Pwr Reserved SCSI_ID 4 SCSI_ID 3 SCSI_ID 2 SCSI_ID 1
I assume the Terminator power supplies the power to the terminator on the last block? This jumper is set to on on all the drives.
Only one drive needs to supply terminator power. It probably does not hurt to have more than one. If your drive doesn't have a jumper for termination resistors then it probably uses ACTUAL resistors. These look like this: http://images.slideplayer.com/25/7891558/slides/slide_24.jpg http://marina.mfarris.com/hard_drives/seagate/st3390n/3655n1.gif http://www.warp12racing.com/images/ataritt/ataritt18.jpg Use Scheme: http://www.craystone.com/dealer/page17.html The thing is, Paul that consumer gear almost always comes terminated because user buy exactly one unit, and can't get it to work if they forget the terminator. Commercial/Professional kit often comes un-terminated because the manufacturers know its going into enclosures which provide termination, and these resistors get thrown away in those cases. But dual use gear tends to come terminated as well. So look for them as physical resistor packs in a socket - not soldered, meant to be removable. The time honored tradition is to yank them from the drives that don't need them and tape them to the top of the drive where they remain till the day you throw it in the trash. -- After all is said and done, more is said than done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (12)
-
Andrei Borzenkov -
Anton Aylward -
Carlos E. R. -
Dave Howorth -
David C. Rankin -
Greg Freemyer -
James Knott -
John Andersen -
John M Andersen
-
Ken Schneider - openSUSE -
Paul Groves -
Wols Lists