[opensuse] dd clone disk resulted in non-bootable system
Hi, I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress For whatever reason system don't start after boot prompt, its seems root partition can't be mounted. I checked in partition manager UUID of new cloned system - it remained the same and listed in /etc/fstab correctly. What I missed there ? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 18/11/17 23:37, andreil1@starlett.lv wrote:
Hi,
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
I checked in partition manager UUID of new cloned system - it remained the same and listed in /etc/fstab correctly.
What I missed there ?
What does "cat /proc/mdstat" give you? What does running "mdadm --examine" or "mdadm --detail" give you? You might find it easier to boot into a rescue system from DVD to explore. I hope it's as simple as mdadm got confused, and you only need to fix it once from inside the rescue CD. Look at the linux raid wiki, and see what you get from that. From what you say, I'm not exactly sure what you've done (I see English is obviously not your first language :-) Does trying to boot off the original disks still work fine? Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sunday, 2017-11-19 at 01:37 +0200, andreil1@starlett.lv wrote:
Hi,
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
Hum. You cloned an installation that was using a single disk to a hardware raid 5 array and expect it to boot? Well, I rather expect this to fail. Actually, I don't see how that could work, but I might be mistaken. I have to think it over slowly.
I checked in partition manager UUID of new cloned system - it remained the same and listed in /etc/fstab correctly.
What I missed there ?
Well... * Not the same MBR, for instance. * 3TB can possibly not be handled with traditional partition, you need gpt. * Different type of boot code. * You need drivers in initrd to handle hw raid perhaps - is it real hw raid, or fake hw raid? - -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAloRDEUACgkQtTMYHG2NR9UArgCfSW/6sb1bDryu+EaKCd2IpqHc /AIAn2MVcJke8uHdrvi9+Pzg1shiV0zQ =NMM3 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
19.11.2017 02:37, andreil1@starlett.lv пишет:
Hi,
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
I checked in partition manager UUID of new cloned system - it remained the same and listed in /etc/fstab correctly.
What I missed there ?
You missed providing even the tiniest bit of factual information that would allow starting to troubleshoot this problem. So each one who answers will silently assume something, (s)he is familiar with, and reply to *that* assumption, not to your actual problem. Start with telling what actually happens when you power on your system What do you see on screen, how far it comes (bootloader, kernel boot, ...), what error messages you get. Of course, telling whether you are using BIOS or EFI is a must for any question about boot problems ... but you won't make it easy, will you? :) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
andreil1@starlett.lv composed on 2017-11-19 01:37 (UTC+0200):
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
I checked in partition manager UUID of new cloned system - it remained the same and listed in /etc/fstab correctly.
What I missed there ?
Before you made the clone, did you include the driver/module for your RAID5 controller in the initrd on the 1TB HD that you cloned from? -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 19 Nov 2017, at 08:30, Felix Miata <mrmazda@earthlink.net> wrote:
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
Before you made the clone, did you include the driver/module for your RAID5 controller in the initrd on the 1TB HD that you cloned from?
Thanks, Felix, this is 90% possibly source of the fault. Is it possible to somehow load 3ware driver at boot and re-generate initrd, either from boot prompt or rescue system ?
-- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation)
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
andreil1 wrote:
On 19 Nov 2017, at 08:30, Felix Miata <mrmazda@earthlink.net> wrote:
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
Before you made the clone, did you include the driver/module for your RAID5 controller in the initrd on the 1TB HD that you cloned from?
Thanks, Felix, this is 90% possibly source of the fault. Is it possible to somehow load 3ware driver at boot and re-generate initrd, either from boot prompt or rescue system ?
Yes, you can explicitly add the module when you call mkinitrd or dracut. With root on the 3ware controller, it should be included automatically when you build the initrd. -- Per Jessen, Zürich (4.8°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/19/2017 12:37 AM, andreil1@starlett.lv wrote:
dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
[...]
What I missed there ?
iflag=fullblock Maybe dd(1) ran into a short read? Have a nice day, Berny -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Hi, I had a problem with non-bootable Leap 42.2 system after dd clone from 1TB disk to hardware RAID 5 (HP P410i with backup battery). Initially, this HD taken from HP DL160 G6 and moved to HP DL360 G7 where it was cloned to RAID5. This 1TB HD boots on HP DL360 G7 without any problem. Thanks to the OpenSuSE community, I'm sure for 90% that reason of non-bootable system is missing HP P410i RAID driver module (hpsa ?) in initrd. This is running server, so I can take it offline for minimum amount of time. Plan is following. Boot from SuSE Argon LiveDVD (which is based on Leap). http://download.opensuse.org/repositories/KDE:/Medias/images/iso/ and follow this instruction written by Knurpht https://forums.opensuse.org/content.php/146-Using-a-LiveCD-to-take-over-repa... su mount /dev/sdX# /mnt Let the installed system be aware of the available hardware mount --bind /dev /mnt/dev Change root to the installed system, mount /proc and /sys to make it complete chroot /mnt mount /proc mount /sys Add "hpsa" to the list of modules in initrd ( /etc/sysconfig/kernel -> INITRD_MODULES="...") and run "mkinitrd". Is this correct or I missing something ? Thanks in advance. On 11/19/2017 01:37 AM, andreil1@starlett.lv wrote:
I have cloned 1TB hard drive to 3TB hardware RAID 5 with dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync status=progress
For whatever reason system don't start after boot prompt, its seems root partition can't be mounted.
I checked in partition manager UUID of new cloned system - it remained the same and listed in /etc/fstab correctly.
What I missed there ?
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
andreil1@starlett.lv wrote:
Hi,
I had a problem with non-bootable Leap 42.2 system after dd clone from 1TB disk to hardware RAID 5 (HP P410i with backup battery). Initially, this HD taken from HP DL160 G6 and moved to HP DL360 G7 where it was cloned to RAID5. This 1TB HD boots on HP DL360 G7 without any problem.
Thanks to the OpenSuSE community, I'm sure for 90% that reason of non-bootable system is missing HP P410i RAID driver module (hpsa ?) in initrd.
Yes, the driver is hpsa.
This is running server, so I can take it offline for minimum amount of time. Plan is following.
Boot from SuSE Argon LiveDVD (which is based on Leap). http://download.opensuse.org/repositories/KDE:/Medias/images/iso/
For this kind of thing, I usually just boot the install system with usessh=1.
mount /dev/sdX# /mnt Let the installed system be aware of the available hardware
My personal favourite, but it does exactly the same: cd /mnt mount --bind /dev dev mount --bind /proc proc mount --bind /sys sys chroot /mnt
Add "hpsa" to the list of modules in initrd ( /etc/sysconfig/kernel -> INITRD_MODULES="...") and run "mkinitrd".
That is most probably not necessary, but it won't hurt. -- Per Jessen, Zürich (4.9°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
andreil1@starlett.lv composed on 2017-11-20 15:40 (UTC+0200):
I had a problem with non-bootable Leap 42.2 system after dd clone from 1TB disk to hardware RAID 5 (HP P410i with backup battery). Initially, this HD taken from HP DL160 G6 and moved to HP DL360 G7 where it was cloned to RAID5. This 1TB HD boots on HP DL360 G7 without any problem.
Thanks to the OpenSuSE community, I'm sure for 90% that reason of non-bootable system is missing HP P410i RAID driver module (hpsa ?) in initrd.
This is running server, so I can take it offline for minimum amount of time. Plan is following.
Boot from SuSE Argon LiveDVD (which is based on Leap). http://download.opensuse.org/repositories/KDE:/Medias/images/iso/
and follow this instruction written by Knurpht https://forums.opensuse.org/content.php/146-Using-a-LiveCD-to-take-over-repa... su mount /dev/sdX# /mnt Let the installed system be aware of the available hardware mount --bind /dev /mnt/dev Change root to the installed system, mount /proc and /sys to make it complete chroot /mnt mount /proc mount /sys
Last comment on that URL indicates a problem with the instruction....
Add "hpsa" to the list of modules in initrd ( /etc/sysconfig/kernel -> INITRD_MODULES="...") and run "mkinitrd".
Is this correct or I missing something ?
According to: https://en.opensuse.org/SDB:Recover_root_password and elsewhere, I chroot thus: mount /dev/sdX# /mnt mount -o bind /proc /mnt/proc mount -o bind /sys /mnt/sys mount -o bind /dev /mnt/dev chroot /mnt Following is alternate syntax: mount /dev/sdX# /mnt mount --bind /proc /mnt/proc mount --bind /sys /mnt/sys mount --bind /dev /mnt/dev chroot /mnt -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Thanks for clarification, Per and Felix ! Is this correct Grub2 reinstall sequence after "chroot /mnt" ? grub2-mkconfig -o /boot/grub2/grub.cfg grub2-install /dev/sdb2 Where sdb2 is a ext4 /boot partition sdb1 is swap, sdb3 - btrf root. On 11/20/2017 04:59 PM, Felix Miata wrote:
andreil1@starlett.lv composed on 2017-11-20 15:40 (UTC+0200):
I had a problem with non-bootable Leap 42.2 system after dd clone from 1TB disk to hardware RAID 5 (HP P410i with backup battery). Initially, this HD taken from HP DL160 G6 and moved to HP DL360 G7 where it was cloned to RAID5. This 1TB HD boots on HP DL360 G7 without any problem. Thanks to the OpenSuSE community, I'm sure for 90% that reason of non-bootable system is missing HP P410i RAID driver module (hpsa ?) in initrd. This is running server, so I can take it offline for minimum amount of time. Plan is following. Boot from SuSE Argon LiveDVD (which is based on Leap). http://download.opensuse.org/repositories/KDE:/Medias/images/iso/ and follow this instruction written by Knurpht https://forums.opensuse.org/content.php/146-Using-a-LiveCD-to-take-over-repa... su mount /dev/sdX# /mnt Let the installed system be aware of the available hardware mount --bind /dev /mnt/dev Change root to the installed system, mount /proc and /sys to make it complete chroot /mnt mount /proc mount /sys Last comment on that URL indicates a problem with the instruction....
Add "hpsa" to the list of modules in initrd ( /etc/sysconfig/kernel -> INITRD_MODULES="...") and run "mkinitrd". Is this correct or I missing something ? According to: https://en.opensuse.org/SDB:Recover_root_password and elsewhere, I chroot thus:
mount /dev/sdX# /mnt mount -o bind /proc /mnt/proc mount -o bind /sys /mnt/sys mount -o bind /dev /mnt/dev chroot /mnt
Following is alternate syntax:
mount /dev/sdX# /mnt mount --bind /proc /mnt/proc mount --bind /sys /mnt/sys mount --bind /dev /mnt/dev chroot /mnt
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/11/17 15:34, andreil1@starlett.lv wrote:
Thanks for clarification, Per and Felix !
Is this correct Grub2 reinstall sequence after "chroot /mnt" ? grub2-mkconfig -o /boot/grub2/grub.cfg grub2-install /dev/sdb2
Where sdb2 is a ext4 /boot partition sdb1 is swap, sdb3 - btrf root.
By the way, something else important :-) Somebody had a glitch on their array because they swapped a disk out, then tried to add it back as a spare except mdadm got confused. Make sure you use a *different* machine (or hotplug this drive *after* your array is up and running), and do a "mdadm --wipe-superblock" on the drive. Otherwise, if this disk gets back into the machine, you could find yourself with a recovery scenario on your hands :-( Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/20/2017 06:32 PM, Wols Lists wrote:
On 20/11/17 15:34, andreil1@starlett.lv wrote:
Thanks for clarification, Per and Felix !
Is this correct Grub2 reinstall sequence after "chroot /mnt" ? grub2-mkconfig -o /boot/grub2/grub.cfg grub2-install /dev/sdb2
Where sdb2 is a ext4 /boot partition sdb1 is swap, sdb3 - btrf root. By the way, something else important :-)
Somebody had a glitch on their array because they swapped a disk out, then tried to add it back as a spare except mdadm got confused.
Make sure you use a *different* machine (or hotplug this drive *after* your array is up and running), and do a "mdadm --wipe-superblock" on the drive.
Otherwise, if this disk gets back into the machine, you could find yourself with a recovery scenario on your hands :-(
Cheers, Wol
Hi, Wol, Thanks. RAID5 will be used on the same machine it is installed right now. BTW, mdadm is related to osftware RAID, in this particular case I have hardrware RAID. I stopped using software RAIDs long time ago. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 20/11/17 20:14, andreil1@starlett.lv wrote:
By the way, something else important :-)
Somebody had a glitch on their array because they swapped a disk out, then tried to add it back as a spare except mdadm got confused.
Make sure you use a *different* machine (or hotplug this drive *after* your array is up and running), and do a "mdadm --wipe-superblock" on the drive.
Otherwise, if this disk gets back into the machine, you could find yourself with a recovery scenario on your hands :-(
Cheers, Wol
Hi, Wol,
Thanks. RAID5 will be used on the same machine it is installed right now. BTW, mdadm is related to osftware RAID, in this particular case I have hardrware RAID. I stopped using software RAIDs long time ago.
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED. Not a good idea. So *please* make sure you wipe all trace of the raid from the *old* disk, before you have an accident ... :-) (As the system boots, mdadm will add drives to the raid as udev finds them. If udev finds the old disk before it finds the disk that replaced it, mdadm will assemble the array with the wrong drives, and the result will be a bit of a mess. Not a problem with raids 5 & 6, but I gather you can trash a mirror pretty easily :-( Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Wols Lists wrote:
On 20/11/17 20:14, andreil1@starlett.lv wrote:
By the way, something else important :-)
Somebody had a glitch on their array because they swapped a disk out, then tried to add it back as a spare except mdadm got confused.
Make sure you use a *different* machine (or hotplug this drive *after* your array is up and running), and do a "mdadm --wipe-superblock" on the drive.
Otherwise, if this disk gets back into the machine, you could find yourself with a recovery scenario on your hands :-(
Cheers, Wol
Hi, Wol,
Thanks. RAID5 will be used on the same machine it is installed right now. BTW, mdadm is related to osftware RAID, in this particular case I have hardrware RAID. I stopped using software RAIDs long time ago.
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED.
Yes, but Andrei is using an HP Smart Array Controller, not software raid. -- Per Jessen, Zürich (8.6°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/21/2017 09:19 AM, Per Jessen wrote:
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED. Yes, but Andrei is using an HP Smart Array Controller, not software raid.
I've also been bitten by hidden RAID metadata on individual disks. I've been using a very simple/dumb script to make sure all the metadata is gone. It blasts the MBR, partition table, and any other cruft at the start of the disk too. Call with the /dev/sdx as the argument. Use at your own risk! #!/bin/bash dd if=/dev/zero of=$1 bs=512 count=2018 dd if=/dev/zero of=$1 bs=512 seek=$(( $(blockdev --getsz $1) - 1024 )) count=1024 exit BTW, I've also seen software raid (mdadm) get confused by disks out of a hardware raid controller. Regards, Lew "When in doubt, zero it out!" -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Lew Wolfgang wrote:
On 11/21/2017 09:19 AM, Per Jessen wrote:
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED. Yes, but Andrei is using an HP Smart Array Controller, not software raid.
I've also been bitten by hidden RAID metadata on individual disks.
oh, me too. BTDT. I don't know how a hardware controller usually deals with it - if a disk isn't actually dead, but maybe kicked out due to timeouts or smart warnings, removing all trace of its RAID "membership" sounds like a sound thing to do when it is failed.
BTW, I've also seen software raid (mdadm) get confused by disks out of a hardware raid controller.
Now that's an interesting one. -- Per Jessen, Zürich (7.1°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/11/17 19:22, Per Jessen wrote:
Lew Wolfgang wrote:
On 11/21/2017 09:19 AM, Per Jessen wrote:
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED. Yes, but Andrei is using an HP Smart Array Controller, not software raid.
I've also been bitten by hidden RAID metadata on individual disks. oh, me too. BTDT.
I don't know how a hardware controller usually deals with it - if a disk isn't actually dead, but maybe kicked out due to timeouts or smart warnings, removing all trace of its RAID "membership" sounds like a sound thing to do when it is failed.
I didn't know this (it should have been obvious, really), but when mdadm kicks a drive out, it updates the superblock to indicate that the drive has failed. Because it's a read, followed by a write, failure that gets a drive kicked, this is pretty safe. If the drive is "iffy", the superblock write will probably succeed. If the superblock write fails, there's a good chance the drive is dead. If you then (re)add this drive back into the array, mdadm checks for a bitmap (or now recently) a journal. If it finds one, it knows which writes have occurred since the drive failure, and just replays the failed writes. (This, if you have a bitmap, is the cause of the "raid 5 write hole", where mdadm can corrupt your array for you :-( If it can't find a bitmap/journal, it just assumes the drive is random data, and rebuilds it from scratch. That's why it's quite dangerous to remove and replace a drive without failing it first - that drive is then a loose cannon that can corrupt your array if it ever gets put back by accident. Which, although I didn't realise it when I wrote it, is a VERY good reason for following the wiki's advice and always using the --replace option to replace a working drive - it safely gets rid of the old drive while copying the data across so you don't lose redundancy while you do so. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/11/17 19:13, Lew Wolfgang wrote:
On 11/21/2017 09:19 AM, Per Jessen wrote:
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED. Yes, but Andrei is using an HP Smart Array Controller, not software raid.
I've also been bitten by hidden RAID metadata on individual disks. I've been using a very simple/dumb script to make sure all the metadata is gone. It blasts the MBR, partition table, and any other cruft at the start of the disk too. Call with the /dev/sdx as the argument. Use at your own risk!
#!/bin/bash dd if=/dev/zero of=$1 bs=512 count=2018 dd if=/dev/zero of=$1 bs=512 seek=$(( $(blockdev --getsz $1) - 1024 )) count=1024 exit
Does the second line wipe the END of the disk? Note that your backup GPT is located at the end of the disk, and a v0.9 or v1.0 raid superblock is also located at the end of the partition/disk. Not knowing where everything is stored, that's why my preference is simply to wipe the entire disk. Snag, of course, is that it takes a LLOOOONNGGG time on today's huge disks.
BTW, I've also seen software raid (mdadm) get confused by disks out of a hardware raid controller.
What's the betting the firmware in said contoller is a cut-down linux/mdadm combo? :-)
Regards, Lew
"When in doubt, zero it out!"
Spot on! Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 El 2017-11-21 a las 20:21 -0000, Wol's lists escribió:
On 21/11/17 19:13, Lew Wolfgang wrote:
On 11/21/2017 09:19 AM, Per Jessen wrote:
Sorry, I don't think I made myself clear. If the old disk was swapped out and copied, then the old disk is "clean". If this disk gets back into the same machine as the raid array, and the machine is rebooted, mdadm IS GOING TO GET CONFUSED. Yes, but Andrei is using an HP Smart Array Controller, not software raid.
I've also been bitten by hidden RAID metadata on individual disks. I've been using a very simple/dumb script to make sure all the metadata is gone. It blasts the MBR, partition table, and any other cruft at the start of the disk too. Call with the /dev/sdx as the argument. Use at your own risk!
#!/bin/bash dd if=/dev/zero of=$1 bs=512 count=2018 dd if=/dev/zero of=$1 bs=512 seek=$(( $(blockdev --getsz $1) - 1024 )) count=1024 exit
Does the second line wipe the END of the disk?
Yes. It erases to the end from 1024 blocks before (of 512 bytes). I think the code should use bigger size block for speed. 1 MB at least. - -- Cheers Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iF4EAREIAAYFAloUjB8ACgkQja8UbcUWM1z78QD/VA2k1c1yTon7GwpfpurLQCza GVBOOjwpyYTaxCNAOW4A/AjfBg3IFrkLH6vY+FFHA2IN7vNsx2EO5HKiulXEJPQe =Lvk4 -----END PGP SIGNATURE-----
On 11/21/2017 12:21 PM, Wol's lists wrote:
#!/bin/bash dd if=/dev/zero of=$1 bs=512 count=2018 dd if=/dev/zero of=$1 bs=512 seek=$(( $(blockdev --getsz $1) - 1024 )) count=1024 exit
Does the second line wipe the END of the disk?
Yes, our hardware RAID (LSI) appends a fairly large number of sectors at the end of the disk for metadata. The size of this area seems to change depending on the card model number, the 1024 blocks seems to cover all the bases. BTW, when the RAID controller reports the size of the disk, it doesn't include this metadata area. So a disk connected to a RAID controller appears to be smaller than a directly connected one. I think mdadm does the same thing.
Note that your backup GPT is located at the end of the disk, and a v0.9 or v1.0 raid superblock is also located at the end of the partition/disk.
Not knowing where everything is stored, that's why my preference is simply to wipe the entire disk. Snag, of course, is that it takes a LLOOOONNGGG time on today's huge disks.
Yes, we're routinely using Seagate 10-TB disks now. They're interesting, the top of the disk is welded into place to try to keep the helium inside. There's a low-helium alarm too. They're warrantied for 5-years, so we'll see.
BTW, I've also seen software raid (mdadm) get confused by disks out of a hardware raid controller.
What's the betting the firmware in said contoller is a cut-down linux/mdadm combo? :-)
There seems to be a common signature element between the LSI controllers and mdadm. Which came first? Regards, Lew -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 El 2017-11-21 a las 12:42 -0800, Lew Wolfgang escribió:
On 11/21/2017 12:21 PM, Wol's lists wrote:
Yes, we're routinely using Seagate 10-TB disks now. They're interesting, the top of the disk is welded into place to try to keep the helium inside. There's a low-helium alarm too. They're warrantied for 5-years, so we'll see.
I guess then that the gas is presurized. There might be some type of reservoir canister as well. Why helium, and not nitrogen? Helium is a noble gas, reacts with nothing. But nitrogen reacts with few things at those temperatures and is cheaper. - -- Cheers Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iF4EAREIAAYFAloUklYACgkQja8UbcUWM1w/nwEAmdEiQEB+3xNZH2HgnFQ8U14L i401niNmURaHEwroi3EA/iUa+0R5Y1pPpHVcIPukpBrZA7fjx4Ai4xC0sngJwC93 =XOyj -----END PGP SIGNATURE-----
On 11/21/2017 12:53 PM, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
El 2017-11-21 a las 12:42 -0800, Lew Wolfgang escribió:
On 11/21/2017 12:21 PM, Wol's lists wrote:
Yes, we're routinely using Seagate 10-TB disks now. They're interesting, the top of the disk is welded into place to try to keep the helium inside. There's a low-helium alarm too. They're warrantied for 5-years, so we'll see.
I guess then that the gas is presurized. There might be some type of reservoir canister as well.
Why helium, and not nitrogen?
Helium is a noble gas, reacts with nothing. But nitrogen reacts with few things at those temperatures and is cheaper.
I "think" it has to do with the size of the atoms/molecules? Nitrogen has huge atoms compared to helium's and they'd interfere with how low the read/write heads can fly above the surface of the spinning media. Head altitude and data density are inversely proportional. Helium, being such a small atom, is hard to confine, which explains welding the case closed. Of course, I may be all wet here... Regards, Lew -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 El 2017-11-21 a las 13:04 -0800, Lew Wolfgang escribió:
On 11/21/2017 12:53 PM, Carlos E. R. wrote:
El 2017-11-21 a las 12:42 -0800, Lew Wolfgang escribió:
On 11/21/2017 12:21 PM, Wol's lists wrote:
Yes, we're routinely using Seagate 10-TB disks now. They're interesting, the top of the disk is welded into place to try to keep the helium inside. There's a low-helium alarm too. They're warrantied for 5-years, so we'll see.
I guess then that the gas is presurized. There might be some type of reservoir canister as well.
Why helium, and not nitrogen?
Helium is a noble gas, reacts with nothing. But nitrogen reacts with few things at those temperatures and is cheaper.
I "think" it has to do with the size of the atoms/molecules? Nitrogen has huge atoms compared to helium's and they'd interfere with how low the read/write heads can fly above the surface of the spinning media. Head altitude and data density are inversely proportional. Helium, being such a small atom, is hard to confine, which explains welding the case closed. Of course, I may be all wet here...
A nitrogen molecule is still ridiculously small compared to the head gap. I think the number of gas molecule layers in that gap can be calculated, but it would not be trivial for me after so many years from high school. I'm not sure the size of the molecule matters in gases for the number per volume. Re Avogadro number, mol size... Helium is indeed hard to confine, but not so hard as hydrogen, though. Any gas is difficult to confine for five years, I think. The length of the junction is very large in comparison to the volume. They would need a screw per centimeter. Perhaps glue would do with other gases and a reservoir canister. It might be marketing, LOL. - -- Cheers Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iF4EAREIAAYFAloUl5kACgkQja8UbcUWM1zC3QEAjEBWnMjRIT5NG04dTygwupO6 19DnQIGF+en7QWWFbjIBAIk8ikhnXSNmSFUhdj0dIkXUndFt2ZTyHGVTsO8Pyjgh =0n/V -----END PGP SIGNATURE-----
On 11/21/2017 01:16 PM, Carlos E. R. wrote:
A nitrogen molecule is still ridiculously small compared to the head gap. I think the number of gas molecule layers in that gap can be calculated, but it would not be trivial for me after so many years from high school. I'm not sure the size of the molecule matters in gases for the number per volume. Re Avogadro number, mol size...
Helium is indeed hard to confine, but not so hard as hydrogen, though. Any gas is difficult to confine for five years, I think. The length of the junction is very large in comparison to the volume. They would need a screw per centimeter. Perhaps glue would do with other gases and a reservoir canister.
It might be marketing, LOL.
Yeah, it's not the atoms. That was dumb of me. Check here: https://www.seagate.com/files/www-content/product-content/enterprise-hdd-fam... "The feasibility of using helium (He) inside of hard disk drives (HDD) has long been the holy grail of rotating magnetic storage. And after many decades of research and development, there is virtually no debate that replacing air with helium inside the magnetic storage enclosure improves areal density capability (lower windage-induced vibration with low-density helium), reduces the power required to spin (lower spindle power from lower-density helium), and limits increases in device temperature (higher thermal conductivity of helium). With the introduction of the 10TB Enterprise Capacity 3.5 HDD, Seagate offers its helium drive technology to a growing cloud-based data center market that is clamoring for unique and robust storage solutions." Note that this isn't shingled magnetic recording (SMR). (I still like the mental image of a disk head rolling through a field of helium atoms...) Regards, Lew -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 El 2017-11-21 a las 13:46 -0800, Lew Wolfgang escribió:
On 11/21/2017 01:16 PM, Carlos E. R. wrote:
A nitrogen molecule is still ridiculously small compared to the head gap. I think the number of gas molecule layers in that gap can be calculated, but it would not be trivial for me after so many years from high school. I'm not sure the size of the molecule matters in gases for the number per volume. Re Avogadro number, mol size...
Helium is indeed hard to confine, but not so hard as hydrogen, though. Any gas is difficult to confine for five years, I think. The length of the junction is very large in comparison to the volume. They would need a screw per centimeter. Perhaps glue would do with other gases and a reservoir canister.
It might be marketing, LOL.
Yeah, it's not the atoms. That was dumb of me. Check here:
https://www.seagate.com/files/www-content/product-content/enterprise-hdd-fam...
"The feasibility of using helium (He) inside of hard disk drives (HDD) has long been the holy grail of rotating magnetic storage. And after many decades of research and development, there is virtually no debate that replacing air with helium inside the magnetic storage enclosure improves areal density capability (lower windage-induced vibration with low-density helium), reduces the power required to spin (lower spindle power from lower-density helium), and limits increases in device temperature (higher thermal conductivity of helium). With the introduction of the 10TB Enterprise Capacity 3.5 HDD, Seagate offers its helium drive technology to a growing cloud-based data center market that is clamoring for unique and robust storage solutions."
Ah, those things make sense. Helium has lower friction than other gases like nitrogen. Still some "gas friction" is needed, or the head would crash with the surface. Like an airplane.
Note that this isn't shingled magnetic recording (SMR).
(I still like the mental image of a disk head rolling through a field of helium atoms...)
:-D - -- Cheers Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iF4EAREIAAYFAloUoTwACgkQja8UbcUWM1yzfgD/T1wqE8b1R2PvdDFrPOok500y NHfRqaZAuuDlqsd6/OYA/1oRBEsWT8gDFK94rhUP7U5lFWAZjyZOEAiH7kEgFeTi =bL7M -----END PGP SIGNATURE-----
On 11/21/2017 01:57 PM, Carlos E. R. wrote:
Ah, those things make sense. Helium has lower friction than other gases like nitrogen.
Still some "gas friction" is needed, or the head would crash with the surface. Like an airplane.
Ah, I don't you need friction to generate aerodynamic lift. An aircraft would still fly through a friction-less atmosphere, thanks to Daniel Bernoulli. Granted, a flying disk head "might" not use Bernoulli, it might maintain its altitude by virtue of the ram effect. (gas rammed into the media/head gap giving it a cushion) This can be noticed in aircraft when they're landing and is called "ground effect". Am I all wet again? Regards, Lew -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/11/17 21:46, Lew Wolfgang wrote:
On 11/21/2017 01:16 PM, Carlos E. R. wrote:
A nitrogen molecule is still ridiculously small compared to the head gap. I think the number of gas molecule layers in that gap can be calculated, but it would not be trivial for me after so many years from high school. I'm not sure the size of the molecule matters in gases for the number per volume. Re Avogadro number, mol size...
Helium is indeed hard to confine, but not so hard as hydrogen, though. Any gas is difficult to confine for five years, I think. The length of the junction is very large in comparison to the volume. They would need a screw per centimeter. Perhaps glue would do with other gases and a reservoir canister.
It might be marketing, LOL.
Yeah, it's not the atoms. That was dumb of me. Check here:
https://www.seagate.com/files/www-content/product-content/enterprise-hdd-fam...
"The feasibility of using helium (He) inside of hard disk drives (HDD) has long been the holy grail of rotating magnetic storage. And after many decades of research and development, there is virtually no debate that replacing air with helium inside the magnetic storage enclosure improves areal density capability (lower windage-induced vibration with low-density helium), reduces the power required to spin (lower spindle power from lower-density helium), and limits increases in device temperature (higher thermal conductivity of helium). With the introduction of the 10TB Enterprise Capacity 3.5 HDD, Seagate offers its helium drive technology to a growing cloud-based data center market that is clamoring for unique and robust storage solutions."
Note that this isn't shingled magnetic recording (SMR).
(I still like the mental image of a disk head rolling through a field of helium atoms...)
Note that (and yes this does sound weird) Helium is actually a metal. And it has all these weird properties that makes it useful for things like this. As mentioned, all gases have the same number of molecules per volume at any given temperature, so replacing Nitrogen (atomic weight 28 per molecule) with Helium (atomic weight 4) abrades less energy off the platter (lower friction :-) and the draught can't disturb the heads so much. The implication is the hydrogen (weight 2) would be even better, but I suspect the fact that helium is a noble gas has something to do with it. As I said, Helium is a metal which is probably why, what heat is generated, helium conducts it away much better. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 El 2017-11-21 a las 22:12 -0000, Wol's lists escribió:
On 21/11/17 21:46, Lew Wolfgang wrote:
On 11/21/2017 01:16 PM, Carlos E. R. wrote:
...
Note that (and yes this does sound weird) Helium is actually a metal.
You can get iron as a gas, too :-)
And it has all these weird properties that makes it useful for things like this. As mentioned, all gases have the same number of molecules per volume at any given temperature, so replacing Nitrogen (atomic weight 28 per molecule) with Helium (atomic weight 4) abrades less energy off the platter (lower friction :-) and the draught can't disturb the heads so much.
They do need some drought, to fly. With no friction (vacuum) the heads would crash.
The implication is the hydrogen (weight 2) would be even better, but I suspect the fact that helium is a noble gas has something to do with it.
It burns. It burns very well, even explode in the right mixture. I know that from personal experience :-) And it is more difficult to contain for 5 years.
As I said, Helium is a metal which is probably why, what heat is generated, helium conducts it away much better.
Maybe. I know nothing about why different gases may conduct heat differently, even if they do. - -- Cheers Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iF4EAREIAAYFAloUrJwACgkQja8UbcUWM1wYGgD/RlK1oIjq/6PoRNe7Vd0LG+5g djuzAjfqiTGUOBP1wsQA/0XDx4CA7Q+rhaYcIX1DVaiwvDC9X7XzpQVlsa0U9oK6 =dcS1 -----END PGP SIGNATURE-----
Hello, On Tue, 21 Nov 2017, Carlos E. R. wrote:
El 2017-11-21 a las 22:12 -0000, Wol's lists escribió:
On 21/11/17 21:46, Lew Wolfgang wrote:
On 11/21/2017 01:16 PM, Carlos E. R. wrote: Note that (and yes this does sound weird) Helium is actually a metal.
You can get iron as a gas, too :-)
And it has all these weird properties that makes it useful for things like this. As mentioned, all gases have the same number of molecules per volume at any given temperature, so replacing Nitrogen (atomic weight 28 per molecule) with Helium (atomic weight 4) abrades less energy off the platter (lower friction :-) and the draught can't disturb the heads so much.
They do need some drought, to fly. With no friction (vacuum) the heads would crash.
Question is: how much? :)
The implication is the hydrogen (weight 2) would be even better, but I suspect the fact that helium is a noble gas has something to do with it.
It burns. It burns very well, even explode in the right mixture. I know that from personal experience :-)
And it is more difficult to contain for 5 years.
As I said, Helium is a metal which is probably why, what heat is generated, helium conducts it away much better.
Maybe. I know nothing about why different gases may conduct heat differently, even if they do.
After some Wikipediing: https://en.wikipedia.org/wiki/Covalent_radius https://en.wikipedia.org/wiki/Van_der_Waals_radius https://en.wikipedia.org/wiki/Molar_volume https://en.wikipedia.org/wiki/Thermal_conductivity cov.r VdW r Molar volume Density Th.cond pm pm 10^~6 m3/mol kg*m^~3 W/(m*K) H: 31±5 120 11,42(solid) 0,0899 0.1805 He: 28 140 21,00(solid) 0,1785 0.1513 N: 71 155 13,54(solid) 1,250 0.02583 From the He-"uses" section: "Helium is used for many purposes that require some of its unique properties, such as its low boiling point, low density, low solubility, high thermal conductivity, or inertness." Couldn't find stuff on molecular H_2. I'd guess a H_2 molecule is larger. Hm. If I read that right, a H_2 molecule has a radius of twice the covalent radius of H, i.e. ~62, i.e. it is larger than that single He atom flying about. And He is just about always single. And H just about never and is "corrosive", cue your good ole NiMH accumulators. And N can get very nasty if not in a pair as N_2[1] and N is about 6 times less thermally conductive as He. He though is AFAIK as inert as it gets. -dnh [1] http://blogs.sciencemag.org/pipeline/archives/2013/01/09/things_i_wont_work_... "The compound exploded in solution, it exploded on any attempts to touch or move the solid, and (most interestingly) it exploded when they were trying to get an infrared spectrum of it." I suppose you could say "if as much as a gnat farts ... in the next room ..." -- After a prolonged soak the parts required to fix the FM could be separated from a rather indescribable, um, something. With a little more resolve it might have walked off and started a certain software company, but maybe it was already too smart for that. -- Rik Steenwinkel, in asr, on the meeting of a Nikon FE and a jar of yoghurt -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Wednesday, 2017-11-22 at 04:12 +0100, David Haller wrote:
Hello,
On Tue, 21 Nov 2017, Carlos E. R. wrote:
They do need some drought, to fly. With no friction (vacuum) the heads would crash.
Question is: how much? :)
Maybe the word "friction" is not appropriate. Friction, aerodynamics, viscosity, whatever :-) Apparently it is enough whatever with helium, or it would not work - the plater linear speed must be high.
The implication is the hydrogen (weight 2) would be even better, but I suspect the fact that helium is a noble gas has something to do with it.
It burns. It burns very well, even explode in the right mixture. I know that from personal experience :-)
And it is more difficult to contain for 5 years.
As I said, Helium is a metal which is probably why, what heat is generated, helium conducts it away much better.
Maybe. I know nothing about why different gases may conduct heat differently, even if they do.
After some Wikipediing:
https://en.wikipedia.org/wiki/Covalent_radius https://en.wikipedia.org/wiki/Van_der_Waals_radius https://en.wikipedia.org/wiki/Molar_volume https://en.wikipedia.org/wiki/Thermal_conductivity
cov.r VdW r Molar volume Density Th.cond pm pm 10^~6 m3/mol kg*m^~3 W/(m*K) H: 31±5 120 11,42(solid) 0,0899 0.1805 He: 28 140 21,00(solid) 0,1785 0.1513 N: 71 155 13,54(solid) 1,250 0.02583
Curious. Nitrogen has a very low thermal conductivity. Both He and H have it higher. The molar volume was the same for all gases, wasn't it? I don't remember.
From the He-"uses" section: "Helium is used for many purposes that require some of its unique properties, such as its low boiling point, low density, low solubility, high thermal conductivity, or inertness."
Couldn't find stuff on molecular H_2. I'd guess a H_2 molecule is larger. Hm. If I read that right, a H_2 molecule has a radius of twice the covalent radius of H, i.e. ~62, i.e. it is larger than that single He atom flying about. And He is just about always single. And H just about never and is "corrosive", cue your good ole NiMH accumulators. And N can get very nasty if not in a pair as N_2[1] and N is about 6 times less thermally conductive as He. He though is AFAIK as inert as it gets.
I forgot that Hidrogen is normally a molecule of two.
-dnh
[1] http://blogs.sciencemag.org/pipeline/archives/2013/01/09/things_i_wont_work_... "The compound exploded in solution, it exploded on any attempts to touch or move the solid, and (most interestingly) it exploded when they were trying to get an infrared spectrum of it." I suppose you could say "if as much as a gnat farts ... in the next room ..."
Nitrogen is used here to fill car tyres by some people (I don't know percent). It is more inert that plain air, although I have my doubts it makes a measurable difference on tyres. I replaced mine yesterday: I made 91000 Km with them (filled with plain air). The mechanic told me the average is 45000. I don't think they would last longer for me if filled with nitrogen. - -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAloVdCYACgkQtTMYHG2NR9UqEQCdH3MMOEEjSNyRwcSXoewoDx1W h5oAmwe9HyGYnbkm2/Jo71vhIPKdBVRX =jaMO -----END PGP SIGNATURE-----
On 11/21/2017 03:16 PM, Carlos E. R. wrote:
Helium is indeed hard to confine, but not so hard as hydrogen,
Oh great... Exploding hard drives... (News @10:00 - Seven dead when server farm explodes... Apparently a hydrogen filled drive ignited causing a chain reaction resulting in a BLEVE that engulfed the 3rd and 4th floors of YourAlwaysOnCloud building -- more at 11:00) -- David C. Rankin, J.D.,P.E.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Wednesday, 2017-11-22 at 02:58 -0600, David C. Rankin wrote:
On 11/21/2017 03:16 PM, Carlos E. R. wrote:
Helium is indeed hard to confine, but not so hard as hydrogen,
Oh great... Exploding hard drives...
(News @10:00 - Seven dead when server farm explodes... Apparently a hydrogen filled drive ignited causing a chain reaction resulting in a BLEVE that engulfed the 3rd and 4th floors of YourAlwaysOnCloud building -- more at 11:00)
ROTFL! :-D Fortunately, the hidrogen that can be contained in the volume of a hard disk enclosure is barely enough to rip it open. :-) Unless they use a small presurized canister for replacing the gas that escapes. - -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAloVbfYACgkQtTMYHG2NR9XcpQCeIFRp8BiX55LCyz7+6zbS9rzb 8PwAn0IIpB2AkhUo5o+ZERn/AlFanTWg =770O -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
* David C. Rankin <drankinatty@suddenlinkmail.com> [11-22-17 04:00]:
On 11/21/2017 03:16 PM, Carlos E. R. wrote:
Helium is indeed hard to confine, but not so hard as hydrogen,
Oh great... Exploding hard drives...
(News @10:00 - Seven dead when server farm explodes... Apparently a hydrogen filled drive ignited causing a chain reaction resulting in a BLEVE that engulfed the 3rd and 4th floors of YourAlwaysOnCloud building -- more at 11:00)
BREAKING News: Server farm explosion now being attributed to NSA. Appears external probing by the National Screwit Agency over-heated a drive ... -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 21/11/17 20:42, Lew Wolfgang wrote:
What's the betting the firmware in said contoller is a cut-down linux/mdadm combo? :-)
There seems to be a common signature element between the LSI controllers and mdadm. Which came first?
Not a clue :-) It wouldn't surprise me if LSI came before mdadm - certainly early linux raid didn't have a superblock (look at the info on the legacy --build option to mdadm :-) Mind you, hardware raid might have made the same assumptions and not had a superblock. But my guess is that the hardware controllers got it first, the linux guys looked at it and said "this is a good thing", and rather than re-invent the wheel, pinched most of it. Thing is, there are several "official" raid standards, which linux can handle, but it prefers to do its own thing. And linux raid has evolved through three or four generations. Gen 1: the --build era. Can't remember the name of the config file, arrays are assembled in user space. Gen 2/3: the version 0 superblock (I only know of v0.9), and the raid array filesystem code type. I don't know how all this lot fits together - it's all obsolete - but the crucial features were that because the fileystem doesn't know about the raid, you can access the filesystem directly without the raid (read-only only of course). Then once the raid is up and running you (re)mount the raid as root read/write. Gen 4: the version 1 superblock. All the version 1 superblocks themselves are identical, it's only the location that changes. 1.0, like 0.9. is at the end of the device so you can access the filesystem outwith the raid and boot off it. Note that the raid partition type is now meaningless, and the kernel doesn't know anything about v1 superblocks, so all raid management MUST be done in userspace. V1.1 stores the raid at the start of the device. This is not usually considered a good idea because, especially if you give a disk rather than a partition to the raid, a lot of things stick their metadata at the start of the device (fdisk, gdisk for example) and it's too easy to trash it by mistake. Both v1.1 and v1.2 reserve a chunk off space at the start of the device, just that v1.2 stores the superblock 4K into device. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Monday, 2017-11-20 at 17:34 +0200, andreil1@starlett.lv wrote:
On 11/20/2017 04:59 PM, Felix Miata wrote:
andreil1@starlett.lv composed on 2017-11-20 15:40 (UTC+0200):
According to: https://en.opensuse.org/SDB:Recover_root_password and elsewhere, I chroot thus:
mount /dev/sdX# /mnt mount -o bind /proc /mnt/proc mount -o bind /sys /mnt/sys mount -o bind /dev /mnt/dev chroot /mnt
Thanks for clarification, Per and Felix !
Is this correct Grub2 reinstall sequence after "chroot /mnt" ? grub2-mkconfig -o /boot/grub2/grub.cfg grub2-install /dev/sdb2
Where sdb2 is a ext4 /boot partition sdb1 is swap, sdb3 - btrf root.
You did not say before that /boot was a separate partition. In this case, you must mount first sdb3 to /mnt, and then sdb2 to /mnt/boot. - -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAloTHy0ACgkQtTMYHG2NR9VTlACffXs+Hwz8TRp/iz7BixfekaAh QrwAn2x5WMQaBReRbkJILno2tqEmRTiX =mf30 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 11/20/2017 08:29 PM, Carlos E. R. wrote:
You did not say before that /boot was a separate partition. In this case, you must mount first sdb3 to /mnt, and then sdb2 to /mnt/boot.
OK, great, thanks ! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Monday, 2017-11-20 at 20:43 +0200, andreil1@starlett.lv wrote:
On 11/20/2017 08:29 PM, Carlos E. R. wrote:
You did not say before that /boot was a separate partition. In this case, you must mount first sdb3 to /mnt, and then sdb2 to /mnt/boot.
OK, great, thanks !
Yes, because grub itself (the boot code) is installed elsewhere, but initrd goes there. If you have other partitions like /usr or /var, you should mount them as well. Things like /home or /data are irrelevant on this scenario. Another trick to get the grub code installed or updated is to start yast in text mode on the chrooted terminal, go to the yast boot module, and there just change the timeout 1 second up or down. This simple change forces YaST to install the entire thing ;-) - -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAloTNO8ACgkQtTMYHG2NR9W+nQCfYbkps6kH+1gTqB5dY1yf0NaX 88AAn3o6NE/2sLeeoawN7IpGa6Qaw9cD =pilo -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (13)
-
Andrei Borzenkov -
andreil1 -
andreil1＠starlett.lv
-
Bernhard Voelker -
Carlos E. R. -
David C. Rankin -
David Haller -
Felix Miata -
Lew Wolfgang -
Patrick Shanahan -
Per Jessen -
Wol's lists -
Wols Lists