[opensuse] ext3 check forced = frustration
Anyone else getting annoyed by the periodic check forced in ext3 partitions? I started using ext3 in my large /home partitions, as recommended by openSUSE (still using Reiser for my smaller RAID1 root partition, because of this strange bug: https://bugzilla.novell.com/show_bug.cgi?id=350992 ) and today I passed again the 60 days limit without checking ext3 and I had to wait 10 minutes for the 200GB partition to be checked. It is bad enough that we don't get any decent feedback in the splash screen ( https://bugzilla.novell.com/show_bug.cgi?id=344271 ), which I can imagine is unsettling to many new users, who will have no clue about what is happening. What I really question is if these lengthy checks are really needed. My Reiser partition was cleared with a simple journal check (isn't that what journaling systems are about?). Add to that the fact that hard disks are becoming bigger and these forced check times can mean one might need to wait 20 minutes or 1/2 hour to get access to the machine again? Obviously this will happen when you have to leave in 10 minutes and you are just about to print out that vital document and your foot hits the reset switch... What is the use of investing a lot of effort into reducing the boot time by precious seconds to give users a better Linux experience, if you have to wait 10 minutes for fsck every time you boot (since Linux allows you to run several months without reboot, until the next kernel patch comes by). There, I vented... -- Carlos FL "It is not worth an intelligent man's time to be in the majority. By definition, there are already enough people to do that." - G. H. Hardy -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions? Maintaining data integrity is precious, automatically repairing errors in the underlying data structure and reducing fragmentation, leading to faster access, but users should at least get the opportunity to determine when these disk maintenance checks are done. How could this be done ?
Thanks for bringing this up, Carlos. Kind regards Philippe --
I started using ext3 in my large /home partitions, as recommended by openSUSE (still using Reiser for my smaller RAID1 root partition, because of this strange bug: https://bugzilla.novell.com/show_bug.cgi?id=350992 ) and today I passed again the 60 days limit without checking ext3 and I had to wait 10 minutes for the 200GB partition to be checked.
It is bad enough that we don't get any decent feedback in the splash screen ( https://bugzilla.novell.com/show_bug.cgi?id=344271 ), which I can imagine is unsettling to many new users, who will have no clue about what is happening.
What I really question is if these lengthy checks are really needed. My Reiser partition was cleared with a simple journal check (isn't that what journaling systems are about?). Add to that the fact that hard disks are becoming bigger and these forced check times can mean one might need to wait 20 minutes or 1/2 hour to get access to the machine again? Obviously this will happen when you have to leave in 10 minutes and you are just about to print out that vital document and your foot hits the reset switch...
What is the use of investing a lot of effort into reducing the boot time by precious seconds to give users a better Linux experience, if you have to wait 10 minutes for fsck every time you boot (since Linux allows you to run several months without reboot, until the next kernel patch comes by).
There, I vented...
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 February 2008 13:25, Philippe Landau wrote:
Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions?
Maintaining data integrity is precious, automatically repairing errors in the underlying data structure and reducing fragmentation, leading to faster access, but users should at least get the opportunity to determine when these disk maintenance checks are done. How could this be done ?
tune2fs See the options -c max-mount-counts -i interval-between-checks[d|m|w] -- Don -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Don Raboud wrote:
On Saturday 16 February 2008 13:25, Philippe Landau wrote:
Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions? Maintaining data integrity is precious, automatically repairing errors in the underlying data structure and reducing fragmentation, leading to faster access, but users should at least get the opportunity to determine when these disk maintenance checks are done. How could this be done ? tune2fs See the options -c max-mount-counts -i interval-between-checks[d|m|w] Great, so this could be integrated in Yast ?
In the mean time, in terminal, enter: su - (becoming root) mount (listing mounted devices like for example /dev/sda1) tune2fs -c 1 /dev/sda1 (to trigger a check at next reboot) then after reboot, again as root: tune2fs -c 0 /dev/sda1 (to disable the check based on times mounted) tune2fs -i 1m /dev/sda1 (to trigger a check every month). A GUI based control panel could enable triggering checks of all disks at once and help setting our own preferences. Could this be implemented in OpenSuse 11.0 ? Kind regards Philippe -- tune2fs -i 1 /dev/sda1 (to trigger a check in 1 day) tune2fs -i 0 /dev/sda1 (to disable check based on time) Please enable either time or reboot times based checks to make sure your data is checked regularly in case you forget to trigger checks manually. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat February 16 2008 14:45:42 Philippe Landau wrote:
Don Raboud wrote:
tune2fs See the options -c max-mount-counts -i interval-between-checks[d|m|w]
Great, so this could be integrated in Yast ?
In the mean time, in terminal, enter: su - (becoming root) mount (listing mounted devices like for example /dev/sda1) tune2fs -c 1 /dev/sda1 (to trigger a check at next reboot) then after reboot, again as root: tune2fs -c 0 /dev/sda1 (to disable the check based on times mounted) tune2fs -i 1m /dev/sda1 (to trigger a check every month).
A GUI based control panel could enable triggering checks of all disks at once and help setting our own preferences. Could this be implemented in OpenSuse 11.0 ?
tune2fs -i 1 /dev/sda1 (to trigger a check in 1 day) tune2fs -i 0 /dev/sda1 (to disable check based on time) Please enable either time or reboot times based checks to make sure your data is checked regularly in case you forget to trigger checks manually.
Thanks, Don and Phillippe! I will start doing that and I think we should submit an enhancement suggestion in bugzilla. Do you care submitting a more structured suggestion for Yast based on your options above, Phillippe? I will vote for it. -- Carlos FL "It is not worth an intelligent man's time to be in the majority. By definition, there are already enough people to do that." - G. H. Hardy -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Feb 16, 2008 12:25 PM, Philippe Landau
Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions? Maintaining data integrity is precious, automatically repairing errors in the underlying data structure and reducing fragmentation, leading to faster access,
Please see the thread on Defragging and why it is not necessary. Also explain why Reiserfs finds no need for this lengthy process before you rush to defend it. Lets face it, ext3 is simply and ugly skin graft over the warts on ext2. It is only standard because it is common and someone else besides novell does the maintenance, which is also true of Windows. -- ----------JSA--------- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
John Andersen wrote:
Also explain why Reiserfs finds no need for this lengthy process before you rush to defend it.
jfs doesn't do the regular checks either. What about xfs? /Per Jessen, Zürich -- http://www.spamchek.com/ - your spam is our business. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2008-02-17 at 11:30 +0100, Per Jessen wrote:
John Andersen wrote:
Also explain why Reiserfs finds no need for this lengthy process before you rush to defend it.
jfs doesn't do the regular checks either. What about xfs?
No, it doesn't, either. I have some xfs partitions and I don't remember seeing it, just a quick check during boot or mount. Now, if the partition is broken... that's a different issue. Xfs partitions can break royally. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuB2etTMYHG2NR9URAnOxAJ9JWrGAtJdh8x2fFWk7Zlxfk3mYqQCff0ln v/PVjgr0rLXJrvTKL41G960= =Z5s9 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2008-02-16 at 13:14 -0700, Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions?
I started using ext3 in my large /home partitions, as recommended by openSUSE (still using Reiser for my smaller RAID1 root partition, because of this strange bug: https://bugzilla.novell.com/show_bug.cgi?id=350992 ) and today I passed again the 60 days limit without checking ext3 and I had to wait 10 minutes for the 200GB partition to be checked.
It is bad enough that we don't get any decent feedback in the splash screen ( https://bugzilla.novell.com/show_bug.cgi?id=344271 ), which I can imagine is unsettling to many new users, who will have no clue about what is happening.
What I really question is if these lengthy checks are really needed.
Yes, they are needed. Unfortunately. What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
My Reiser partition was cleared with a simple journal check (isn't that what journaling systems are about?).
Yes... but it happens that that fast check doesn't catch all problems. Sometimes user complain of "funny" things happening on their reiserfs partition; when they force a reiserfsck they do discover they had undetected problems.
Add to that the fact that hard disks are becoming bigger and these forced check times can mean one might need to wait 20 minutes or 1/2 hour to get access to the machine again? Obviously this will happen when you have to leave in 10 minutes and you are just about to print out that vital document and your foot hits the reset switch...
Perhaps a method could be devised to warn in advance that a disk check is getting near, and then do it at your convenience. Or when the time comes to have the possibility of saying "tomorrow, please". I read somewhere that the kernel people had a power failure, and it forced an fsck of their filesystem, with all those things they have. It took hours, if I remember correctly. The comment was that a recovery from backup would have been faster. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHt0m8tTMYHG2NR9URAsrsAKCUuybyWHwnxlGfoe8PiOm7ZV+oWwCfTuHk JsJjyW//bEKT3OCa6BIt46A= =gKcH -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 February 2008 03:38:11 pm Carlos E. R. wrote:
The Saturday 2008-02-16 at 13:14 -0700, Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions?
I started using ext3 in my large /home partitions, as recommended by openSUSE (still using Reiser for my smaller RAID1 root partition, because of this strange bug: https://bugzilla.novell.com/show_bug.cgi?id=350992 ) and today I passed again the 60 days limit without checking ext3 and I had to wait 10 minutes for the 200GB partition to be checked.
It is bad enough that we don't get any decent feedback in the splash screen ( https://bugzilla.novell.com/show_bug.cgi?id=344271 ), which I can imagine is unsettling to many new users, who will have no clue about what is happening.
What I really question is if these lengthy checks are really needed.
Yes, they are needed. Unfortunately.
What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
My Reiser partition was cleared with a simple journal check (isn't that what journaling systems are about?).
Yes... but it happens that that fast check doesn't catch all problems. Sometimes user complain of "funny" things happening on their reiserfs partition; when they force a reiserfsck they do discover they had undetected problems.
Add to that the fact that hard disks are becoming bigger and these forced check times can mean one might need to wait 20 minutes or 1/2 hour to get access to the machine again? Obviously this will happen when you have to leave in 10 minutes and you are just about to print out that vital document and your foot hits the reset switch...
Perhaps a method could be devised to warn in advance that a disk check is getting near, and then do it at your convenience. Or when the time comes to have the possibility of saying "tomorrow, please".
I read somewhere that the kernel people had a power failure, and it forced an fsck of their filesystem, with all those things they have. It took hours, if I remember correctly. The comment was that a recovery from backup would have been faster.
-- Cheers, Carlos E. R.
Can the check be ran manually like you can with reiser? That way instead of doing it a boot every 60 days it can be ran while the system is idle.
On 02/17/2008 11:04 AM, Adam Jimerson wrote:
Can the check be ran manually like you can with reiser? That way instead of doing it a boot every 60 days it can be ran while the system is idle.
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO. -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Joe Morris wrote:
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO.
This conversation has me confused, but then, I'm relatively new to Linux. But I've read more than one post claiming that someone's Linux system has been up and running nonstop for nearly three years. What kind of file system do _those_ folks use? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Jerry Houston wrote:
Joe Morris wrote:
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO.
This conversation has me confused, but then, I'm relatively new to Linux.
But I've read more than one post claiming that someone's Linux system has been up and running nonstop for nearly three years. What kind of file system do _those_ folks use?
That was probably me - I posted an uptime from a server which has been up for near 1040 days now. We also have a few dozen linux servers with only around 550 days uptime (more recent kernel update), but none of the servers has ever gone down except for hardware maintenance, a kernel update, or a power failure. They all use reiser, which is BTW the default on suse enterprise. Joe -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO.
That makes sense on a server. You rarely restart a server, and it runs 24x7. Having a forced file system check on a once or twice per year restart is generally no big deal as the server probably down for scheduled maintenance anyway. What about the people who install openSUSE on a laptop? or on a home PC that is shut down every night? Are we right in expecting them to do tune2fs? An experienced user might not mind, but what about the new users? I understand why Novell chose to set ext3 as the default file system, and I know that if I prefer I can pick any other one from the list when I do my custom partitioning.... but where does that leave our new users... people new to Linux? ...people who have no clue about ext3, Reiser, etc. Just some questions that really need to be thought about for a future release of openSUSE... I don't know the "right" answer here... but unleashing ext3 and its lengthy forced fsck on unsuspecting new users is not exactly a good thing. As others have noted, there really needs to be some warning... and some control over the ext3 fsck on boot... something an average new user can deal with. (maybe this is something to be discussed on the Project mailing list?) C. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun February 17 2008 00:25:02 Clayton wrote:
What about the people who install openSUSE on a laptop? or on a home PC that is shut down every night? Are we right in expecting them to do tune2fs? An experienced user might not mind, but what about the new users? I understand why Novell chose to set ext3 as the default file system, and I know that if I prefer I can pick any other one from the list when I do my custom partitioning.... but where does that leave our new users... people new to Linux? ...people who have no clue about ext3, Reiser, etc.
That is exactly what I think, Clayton.
Just some questions that really need to be thought about for a future release of openSUSE... I don't know the "right" answer here... but unleashing ext3 and its lengthy forced fsck on unsuspecting new users is not exactly a good thing. As others have noted, there really needs to be some warning... and some control over the ext3 fsck on boot... something an average new user can deal with. (maybe this is something to be discussed on the Project mailing list?)
You may be right that Project or Factory are more appropriate mailing-lists for this discussion. Feel free to move it over there, but I am not subscribed to them atm. I will carry on through the one or two bugzilla enhancement entries (the one for the boot process - bug 344271 - and a new one for Yast, which Phillippe may create, or else I will). -- Carlos FL "It is not worth an intelligent man's time to be in the majority. By definition, there are already enough people to do that." - G. H. Hardy -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Clayton wrote:
convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO.
That makes sense on a server. You rarely restart a server, and it runs 24x7. Having a forced file system check on a once or twice per year restart is generally no big deal as the server probably down for scheduled maintenance anyway.
What about the people who install openSUSE on a laptop? or on a home PC that is shut down every night? Are we right in expecting them to do tune2fs? An experienced user might not mind, but what about the new users? I understand why Novell chose to set ext3 as the default file system, and I know that if I prefer I can pick any other one from the list when I do my custom partitioning.... but where does that leave our new users... people new to Linux? ...people who have no clue about ext3, Reiser, etc.
Are laptops less susceptible to filesystem corruption?
Just some questions that really need to be thought about for a future release of openSUSE... I don't know the "right" answer here... but unleashing ext3 and its lengthy forced fsck on unsuspecting new users is not exactly a good thing. As others have noted, there really needs to be some warning... and some control over the ext3 fsck on boot... something an average new user can deal with. (maybe this is something to be discussed on the Project mailing list?)
It's never been a problem for me. But then again, I don't put everything into one huge 80 GB filesystem. Several smaller filesystesm will fsck faster than the same amount of disk space in one large partition. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
What about the people who install openSUSE on a laptop? or on a home PC that is shut down every night? Are we right in expecting them to do tune2fs? An experienced user might not mind, but what about the new users? I understand why Novell chose to set ext3 as the default file system, and I know that if I prefer I can pick any other one from the list when I do my custom partitioning.... but where does that leave our new users... people new to Linux? ...people who have no clue about ext3, Reiser, etc.
Are laptops less susceptible to filesystem corruption?
:-) certainly not. Possibly more so due to the more thumps and bumps a laptop has to endure. I named laptops simply because they are typically booted up/rebooted much more often than your typical server (which can go months or even years without a restart). With regular restarts... possibly even several per day, a laptop user is going to bump into the fsck after X number of boots thing a lot sooner. If this person is a new user.. new to Linux, they are going to be rather annoyed and frustrated (personal experience through a very heated phone call from a friend who had exactly this problem/experience with his laptop). Usually (in my experience) a laptop user needs a fast boot up... openSUSE is reasonable most of the time... until ext3 does its fsck.
It's never been a problem for me. But then again, I don't put everything into one huge 80 GB filesystem. Several smaller filesystesm will fsck faster than the same amount of disk space in one large partition.
Right, but you are not an inexperienced user. ext3 is not a problem for those of us who are long time Linux users... we know about ext3's forced fsck. We can plan for it. We know how to partition our drives so as to reduce the impact of the fsck. We know about the other file systems and can and often do select other ones when the job requires it. We can go tinker with tune2fs until we are happy with the setup.... But... I am not talking about the need to solve an issue for experienced users. I am talking about the new users... people new to Linux. openSUSE defaults (with a well discussed reason) to ext3. That is fine. But ext3 has a "gotcha" in the form of the mandatory fsck. This is not currently handled nicely. It is a bit of a surprise to new users when their computer running a _default_ openSUSE suddenly decides that this particular boot is the time to do a check. There is at least one enhancement bug open on this (discussed elsewhere in this thread)... maybe we can improve things a bit for the new users :-) C -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 10:03 +0100, Clayton wrote:
bump into the fsck after X number of boots thing a lot sooner. If this person is a new user.. new to Linux, they are going to be rather annoyed and frustrated (personal experience through a very heated phone call from a friend who had exactly this problem/experience with his laptop). Usually (in my experience) a laptop user needs a fast boot up... openSUSE is reasonable most of the time... until ext3 does its fsck.
Tell them to use hybernation: the computer awakes almost in seconds, ready to use, and no fsck. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuXjftTMYHG2NR9URAqAgAJ9OBJJ6Z+6VDWzHZMW12vVDYqHLrgCeOTp+ h9kHmhtVXzJORi48Ri9a/xs= =h9mh -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
The Monday 2008-02-18 at 10:03 +0100, Clayton wrote:
bump into the fsck after X number of boots thing a lot sooner. If this person is a new user.. new to Linux, they are going to be rather annoyed and frustrated (personal experience through a very heated phone call from a friend who had exactly this problem/experience with his laptop). Usually (in my experience) a laptop user needs a fast boot up... openSUSE is reasonable most of the time... until ext3 does its fsck.
Tell them to use hybernation: the computer awakes almost in seconds, ready to use, and no fsck.
-- Cheers, Carlos E. R.
Works (if openSuSE powersaved supports hybernation for that model) and if the laptop is _not_ moved between different networks while it is asleep. Hostnames and resolv.conf can become a problem in that case. -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 23 February 2008 10:18:15 am David C. Rankin wrote:
Works (if openSuSE powersaved supports hybernation for that model) and if the laptop is _not_ moved between different networks while it is asleep. Hostnames and resolv.conf can become a problem in that case.
The default suspend framework will shutdown / restart the network as part of the hibernate/resume cycle, so there should be no problems there. I suspend my lappie all the time when moving from home wifi to office wifi, it never hiccups. Of course, getting suspend/hibernate to work in the first place is usually the real issue... ;) Cheers, KV -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2008-02-23 at 14:15 -0500, Kevin Valko wrote:
On Saturday 23 February 2008 10:18:15 am David C. Rankin wrote:
Works (if openSuSE powersaved supports hybernation for that model) and if the laptop is _not_ moved between different networks while it is asleep. Hostnames and resolv.conf can become a problem in that case.
Mine is a desktop, no moving around.
The default suspend framework will shutdown / restart the network as part of the hibernate/resume cycle, so there should be no problems there.
I had to add a script to restart the ntp daemon.
I suspend my lappie all the time when moving from home wifi to office wifi, it never hiccups.
If you have it set to use dhcp, it should ask for a new lease on awakening.
Of course, getting suspend/hibernate to work in the first place is usually the real issue... ;)
Indeed! But once it works, it is wonderfull. Saves me whole minutes. I can even bother to hybernate when I go for a coffee! Ok, make it a "long" coffe ;-) - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHwU7atTMYHG2NR9URAnogAJ9cZv+DxHbeYgXfyxzF1IwR1ys5fgCfT7jt DZ16IvxMx0LskpjSIyRcUFE= =vnZB -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Clayton wrote:
Are laptops less susceptible to filesystem corruption?
:-) certainly not. Possibly more so due to the more thumps and bumps a laptop has to endure. ????
If a laptop received such a bump during writing, the result would be a disk error, because the written data isn't where it's supposed to be, fragmented or not. That is not the same thing as fragmentation, where by file system design, data simply gets written in the next convenient spot, whether it's large enough to hold the file or not. If not big enough, the overflow goes to the next convenient spot etc. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
James Knott wrote:
Clayton wrote:
Are laptops less susceptible to filesystem corruption?
:-) certainly not. Possibly more so due to the more thumps and bumps a laptop has to endure. ????
If a laptop received such a bump during writing, the result would be a disk error, because the written data isn't where it's supposed to be, fragmented or not. That is not the same thing as fragmentation, where
This thread is about fsck, and therefore, filesystem corruption. Fragmentation is a different thread.
by file system design, data simply gets written in the next convenient spot, whether it's large enough to hold the file or not. If not big enough, the overflow goes to the next convenient spot etc.
What does that have to do with the need for laptops to be fsck'ed? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Joe Sloan wrote:
Jerry Houston wrote:
Joe Morris wrote:
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO. This conversation has me confused, but then, I'm relatively new to Linux.
But I've read more than one post claiming that someone's Linux system has been up and running nonstop for nearly three years. What kind of file system do _those_ folks use?
That was probably me - I posted an uptime from a server which has been up for near 1040 days now. We also have a few dozen linux servers with only around 550 days uptime (more recent kernel update), but none of the servers has ever gone down except for hardware maintenance, a kernel update, or a power failure.
They all use reiser, which is BTW the default on suse enterprise.
Unfortunately, reiserfs burned me in a power failure situation. In contrast, my laptop suffered a couple power failures (generator maintenance on base which was typically 4+ hours), and with XFS, I have suffered no problems. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2008-02-17 at 18:18 -0500, Aaron Kulkis wrote:
Unfortunately, reiserfs burned me in a power failure situation.
In contrast, my laptop suffered a couple power failures (generator maintenance on base which was typically 4+ hours), and with XFS, I have suffered no problems.
I have been burned by reiserfs, xfs, and ext3. My xfs problem was so bad that the recovery program crashed. That was with opensuse 10.1, 10.2, 10.3, and finally, it was solved a month or two ago - remember an xfsprogs upgrade to 2.9.4-17.1 at the end of January? That was it. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuXpjtTMYHG2NR9URAqFtAJ9JdTtziLsT+BHBWaloA4ZjIjH6JgCgjV8x fMsKgRRpt7iq978XcaRasu8= =34Kz -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Aaron Kulkis wrote:
Joe Sloan wrote:
Jerry Houston wrote:
Joe Morris wrote:
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO. This conversation has me confused, but then, I'm relatively new to Linux. But I've read more than one post claiming that someone's Linux system has been up and running nonstop for nearly three years. What kind of file system do _those_ folks use?
That was probably me - I posted an uptime from a server which has been up for near 1040 days now. We also have a few dozen linux servers with only around 550 days uptime (more recent kernel update), but none of the servers has ever gone down except for hardware maintenance, a kernel update, or a power failure.
They all use reiser, which is BTW the default on suse enterprise.
Unfortunately, reiserfs burned me in a power failure situation.
What distro, how many years ago, and are you sure the hardware wasn't at fault? I've heard of reiser problems in early versions, and especially on redhat. Also, you may have assumed that the filesystem caused your problem when it was simple hardware failure. Joe -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 08:40 -0800, Joe Sloan wrote:
Unfortunately, reiserfs burned me in a power failure situation.
What distro, how many years ago, and are you sure the hardware wasn't at fault?
A software bug destroyed a reiserfs partition I had, in SuSe 9.1, year 2004. The issue is on record in one of the suse lists. Here: http://lists.opensuse.org/opensuse-security/2004-08/msg00234.html - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHubo6tTMYHG2NR9URAn48AKCRfozmh4iqIXIij2OP73CeRAAXcACfXvEy ZkQdKKQPelUxwIOsxQmBUFA= =0NiS -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
The Monday 2008-02-18 at 08:40 -0800, Joe Sloan wrote:
Unfortunately, reiserfs burned me in a power failure situation.
What distro, how many years ago, and are you sure the hardware wasn't at fault?
A software bug destroyed a reiserfs partition I had, in SuSe 9.1, year 2004. The issue is on record in one of the suse lists. Here:
http://lists.opensuse.org/opensuse-security/2004-08/msg00234.html
Interesting - so a hashing bug was introduced when the xattr stuff was patched into the reiser3 kernel code? Interesting... code changes always carry some risk, but I seem to have dodged that particular bullet. However, they diagnosed and squashed that particular kernel bug, and it hasn't reappeared, so that would not really be a factor in whether to use reiserfs now - especially since there have been bugs affecting all of the filesystems at one time or the other. Joe -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 09:23 -0800, Joe Sloan wrote:
Interesting - so a hashing bug was introduced when the xattr stuff was patched into the reiser3 kernel code? Interesting... code changes always carry some risk, but I seem to have dodged that particular bullet.
I was hit because I tried to test it, I was curious.
However, they diagnosed and squashed that particular kernel bug, and it hasn't reappeared, so that would not really be a factor in whether to use reiserfs now - especially since there have been bugs affecting all of the filesystems at one time or the other.
Absolutely. I use reiserfs myself a lot - though rarely as the root partition. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHucKKtTMYHG2NR9URAkU8AJ9x0tu/stlA8A4SaEx097O4zhD/QQCfelfc jsiz/kvetfxFpDMIrwNplsA= =xNGS -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Joe Sloan wrote:
Aaron Kulkis wrote:
Joe Sloan wrote:
Jerry Houston wrote:
Joe Morris wrote:
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO. This conversation has me confused, but then, I'm relatively new to Linux. But I've read more than one post claiming that someone's Linux system has been up and running nonstop for nearly three years. What kind of file system do _those_ folks use? That was probably me - I posted an uptime from a server which has been up for near 1040 days now. We also have a few dozen linux servers with only around 550 days uptime (more recent kernel update), but none of the servers has ever gone down except for hardware maintenance, a kernel update, or a power failure.
They all use reiser, which is BTW the default on suse enterprise. Unfortunately, reiserfs burned me in a power failure situation.
What distro, how many years ago,
About three years ago. 9.x release.
and are you sure the hardware wasn't at fault?
What part of POWER FAILURE did you not understand? The entire neighborhood had no power for a couple days. I have a UPS, but not THAT good. Unfortunately, when it occurred, I was a couple hundred miles away -- well, actually it was fortunate for me personally, because the we never lost power at the military reservation where I was at during this time.
I've heard of reiser problems in early versions, and especially on redhat. Also, you may have assumed that the filesystem caused your problem when it was simple hardware failure.
This would have been reiser 3, I believe.
Joe
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Jerry Houston wrote:
Joe Morris wrote:
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO.
This conversation has me confused, but then, I'm relatively new to Linux.
But I've read more than one post claiming that someone's Linux system has been up and running nonstop for nearly three years. What kind of file system do _those_ folks use?
Take your pick. I've used ext2 & 3 and Reiser. It's not the file system that makes them reliable, it's the operating system. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 February 2008 10:46:03 pm Joe Morris wrote:
On 02/17/2008 11:04 AM, Adam Jimerson wrote:
Can the check be ran manually like you can with reiser? That way instead of doing it a boot every 60 days it can be ran while the system is idle.
Yes and no. Of course you can run it as a cli command, BUT NOT on a mounted rw filesystem. That is why the check on boot, before it is mounted rw. You could remount the filesystem ro, but not very convenient on your root partition. I have decided it is a small price to pay for data consistency for as often as I boot the server. FS corruption left uncorrected is not an advantage in the few minutes it adds to a boot every few months IMO.
-- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64
Yea doing that to my root partition would not make sense, but I should have added that it would be for /home and things like external HDDs that was formatted to ext3. From what I have seen from both my laptop and desktop, both have a 120 GB SATA HDD, the root partition is the smallest that gets checked and takes about 30 seconds to do. No problem for me its just 30 seconds more, so it feels like openSUSE 10.2's boot time but that is no problem. The problem is my /home partition takes a while on both machines to run so I figured if I can run it while they are idle then when my systems it the 60 it would only have to scan my root. I'm also planning on getting an external HDD and mount it on my laptop for backups, I use this more than my desktop, and I have yet to decided on a file format for it, but if I do go with ext3 then I can also just check that by hand. How can the check be done, is it part of tune2fs, or is it a different command?
On 02/18/2008 03:12 AM, Adam Jimerson wrote:
Yea doing that to my root partition would not make sense, but I should have added that it would be for /home and things like external HDDs that was formatted to ext3. That is much less of a problem, especially for a system that is on most of the time. You could easily create a script that ran from cron that umount /home fsck.ext3 -C -p /dev/(partition for home) mount /home or something a bit more elaborate. From what I have seen from both my laptop and desktop, both have a 120 GB SATA HDD, the root partition is the smallest that gets checked and takes about 30 seconds to do. No problem for me its just 30 seconds more, so it feels like openSUSE 10.2's boot time but that is no problem. The problem is my /home partition takes a while on both machines to run so I figured if I can run it while they are idle then when my systems it the 60 it would only have to scan my root. Understood. That idea sounds good, and pretty easily done. I'm also planning on getting an external HDD and mount it on my laptop for backups, I use this more than my desktop, and I have yet to decided on a file format for it, but if I do go with ext3 then I can also just check that by hand.
I do this, and have the file system check as part of the backup script. Since it isn't always connected, I think to make it a part of the backup script makes more sense. Mine also umounts it (wherever it was mounted), checks it, then mounts it where I want it.
How can the check be done, is it part of tune2fs, or is it a different command?
The command depends on the file system. -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2008-02-16 at 22:04 -0500, Adam Jimerson wrote:
On Saturday 16 February 2008 03:38:11 pm Carlos E. R. wrote:
Can the check be ran manually like you can with reiser? That way instead of doing it a boot every 60 days it can be ran while the system is idle.
If I recall, the partitions can not be mounted during the check. Otherwise, I think they would have already established a cron job instead of checking at boot time... I can see this as an argument supporting multiple small partitions for the system (as I do have), and then large "work" partitions for /srv, /local, /data, "/whatever". These work parts could be cron-scripted to be periodically umount/checked (ie monthly), while leaving the remaining system partitions at the mercy of the normal boot-count-trigger mechanism. Would the fsck program be intelligent enough to only check the out-of-range partitions...and thus provide a less-painful wait when it happens? I don't know...(thus this is a question...) ((I still use reiserfs)) Tom in NM -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 02/17/2008 12:16 PM, Tom Patton wrote:
I can see this as an argument supporting multiple small partitions for the system (as I do have), and then large "work" partitions for /srv, /local, /data, "/whatever". True, this would speed up fsck checks, as the partitions would be smaller. Would the fsck program be intelligent enough to only check the out-of-range partitions...and thus provide a less-painful wait when it happens? I don't know...(thus this is a question...)
Yes -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2008-02-17 at 12:46 +0800, Joe Morris wrote:
On 02/17/2008 12:16 PM, Tom Patton wrote:
I can see this as an argument supporting multiple small partitions for the system (as I do have), and then large "work" partitions for /srv, /local, /data, "/whatever". True, this would speed up fsck checks, as the partitions would be smaller.
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuBxMtTMYHG2NR9URAmASAJ9fM+JElpdhEDA/3zCgC4MQ0qACgwCffdiK ok76aE0AcWi2Getjp1vF9H8= =uVdr -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon.
That's a SATA or PATA limit isn't it? /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
* Per Jessen
Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon.
That's a SATA or PATA limit isn't it?
not really. It's a scsi limit. Seems the *ata interface ?translation? is being dropped in favor of scsi which has the 16 (15) device limitation. Rumor that work was in progress to change that limit, but have seen no recent mention :^(. -- Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://counter.li.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Patrick Shanahan wrote:
* Per Jessen
[02-17-08 08:30]: Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon.
That's a SATA or PATA limit isn't it?
not really. It's a scsi limit. Seems the *ata interface ?translation? is being dropped in favor of scsi which has the 16 (15) device limitation. Rumor that work was in progress to change that limit, but have seen no recent mention :^(.
Ah, got it - I guess that limitation is mostly a problem for large ATA drives. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
* Per Jessen
[02-17-08 08:30]: Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon.
That's a SATA or PATA limit isn't it?
not really. It's a scsi limit. Seems the *ata interface ?translation? is being dropped in favor of scsi which has the 16 (15) device limitation. Rumor that work was in progress to change that limit, but have seen no recent mention :^(. -- Yeah, when they made the switch there were a bunch of us screaming about the
On Sunday 17 February 2008 08:35:57 am Patrick Shanahan wrote: partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work! Has anybody heard anything?? Somebody really working on it? I'll bet not, and we will all be left stuck with this "improvement". Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Bob S wrote:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
Has anybody heard anything?? Somebody really working on it? I'll bet not, and we will all be left stuck with this "improvement".
While in some ways I agree with your sentiments, the question I have is what would you do if that was a SCSI drive? Wouldn't that limit still apply regardless? You might consider using LVM as, IIRC, it's not affected by that limit. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 08:15 -0500, James Knott wrote:
While in some ways I agree with your sentiments, the question I have is what would you do if that was a SCSI drive? Wouldn't that limit still apply regardless?
Typically, scsi drives were smaller, and it was easier to add many drives (no limit of 4 drives per system).
You might consider using LVM as, IIRC, it's not affected by that limit.
But it forces to repartition the entire drive, and I don't feel LVM is as safe. One of the mains reasons of having many partition is damage contention, like the partitions on a ship. If you have a bad failure, the damage, or the worst damage, is limited to a partition. I have had my day saved more than once when I had a partition fully and irrecoverably lost. If I had everything on a single partition, I would have lost everything! My fear is that the LVM layer may remove that security. A problem in that layer could damage all partitions inside. I would need assurance, real assurance, that this is not the case. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuYj5tTMYHG2NR9URAk2CAJ0WI6TKIUzSkJR4U/27OmK6qJr44QCgl7xb IzKwnl6TbS4BznLkvl7W0f4= =lzkC -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
One of the mains reasons of having many partition is damage contention, like the partitions on a ship. If you have a bad failure, the damage, or the worst damage, is limited to a partition. I have had my day saved more than once when I had a partition fully and irrecoverably lost. If I had everything on a single partition, I would have lost everything!
I have never heard that reasoning before, but I guess you could be right - nonetheless, if I was worried about disk-failure, I'd buy a second drive and set up a RAID1. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 15:15 +0100, Per Jessen wrote:
Carlos E. R. wrote:
One of the mains reasons of having many partition is damage contention, like the partitions on a ship. If you have a bad failure, the damage, or the worst damage, is limited to a partition. I have had my day saved more than once when I had a partition fully and irrecoverably lost. If I had everything on a single partition, I would have lost everything!
I have never heard that reasoning before, but I guess you could be right - nonetheless, if I was worried about disk-failure, I'd buy a second drive and set up a RAID1.
Raid is no protection at all in this case, because a software problem affects the filesystem structure imposed over the raid. Both copies would be simultaneously broken. Raid only protects against hardware problems. I have been bitten by this kind of failure several times, with every type of filesystem. The only real protection is backup. Having several partitions, or better, several disks, improves it. These kind of failure can be triggered by a hardware failure, like a power outage at a very inopportune moment, followed by another one during recovery. Or a faulty IDE cable failing also when less convenient. Or ram memory corruption (either physical, or another program overwriting it). The software can not cope, the main inodes (or fat) get corrupted, and the structure of the partition it was writing to gets corrupted beyond recovery. Once I had a reiser partition destroyed because of a software bug. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuZaftTMYHG2NR9URAnQ5AJ9bWwtv9z1DxoM6BwxAB6ubfupAzACfTf+1 EHQxl9TDKC6Mvf8kwLzEFUc= =g/Lv -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Carlos E. R. wrote:
One of the mains reasons of having many partition is damage contention, like the partitions on a ship. If you have a bad failure, the damage, or the worst damage, is limited to a partition. I have had my day saved more than once when I had a partition fully and irrecoverably lost. If I had everything on a single partition, I would have lost everything!
I have never heard that reasoning before, but I guess you could be right - nonetheless, if I was worried about disk-failure, I'd buy a second drive and set up a RAID1.
It definitely is an issue. There's not much worse than having root filesystem corruption destroy your LVM table. Now a simple OS-reinstall has turned into a complete restoration of the entire system from your most recent back-up. Been there -- it's not fun. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Monday 2008-02-18 at 08:15 -0500, James Knott wrote:
While in some ways I agree with your sentiments, the question I have is what would you do if that was a SCSI drive? Wouldn't that limit still apply regardless?
Typically, scsi drives were smaller, and it was easier to add many drives (no limit of 4 drives per system).
You might consider using LVM as, IIRC, it's not affected by that limit.
But it forces to repartition the entire drive, and I don't feel LVM is as safe.
It's not, which is why I don't use it. All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
One of the mains reasons of having many partition is damage contention, like the partitions on a ship. If you have a bad failure, the damage, or the worst damage, is limited to a partition. I have had my day saved more than once when I had a partition fully and irrecoverably lost. If I had everything on a single partition, I would have lost everything!
My fear is that the LVM layer may remove that security. A problem in that layer could damage all partitions inside.
I've seen it happen. Not pretty. And definitely not advisable for any machine which is not backed up on a nightly basis.
I would need assurance, real assurance, that this is not the case.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Aaron Kulkis wrote:
But it forces to repartition the entire drive, and I don't feel LVM is as safe.
It's not, which is why I don't use it.
All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
Any other suggestions for something that will let me expand my filesystem on my systems that must be kept running 365x24? LVM and EVMS are both actually used quite a lot because they solve a real problem. Similar systems have been in use in e.g. AIX and HPUX for over 10 years. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 09:29 +0100, Per Jessen wrote:
Aaron Kulkis wrote:
But it forces to repartition the entire drive, and I don't feel LVM is as safe.
It's not, which is why I don't use it.
All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
Any other suggestions for something that will let me expand my filesystem on my systems that must be kept running 365x24?
LVM and EVMS are both actually used quite a lot because they solve a real problem. Similar systems have been in use in e.g. AIX and HPUX for over 10 years.
That's true, but it is not a solution for those of us having twenty physical partitions already in a drive. Every engineering solution has pros and cons. LVM is good for something, physical partitions is good for another. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHurQUtTMYHG2NR9URAunXAJsETuzofItQypc98YFjoBWh+HZ/dQCdGbw9 S0+ow/uKpnqerlyw0kVi9g0= =2RAb -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
But it forces to repartition the entire drive, and I don't feel LVM is as safe.
It's not, which is why I don't use it.
All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
Any other suggestions for something that will let me expand my filesystem on my systems that must be kept running 365x24?
LVM and EVMS are both actually used quite a lot because they solve a real problem. Similar systems have been in use in e.g. AIX and HPUX for over 10 years.
That's true, but it is not a solution for those of us having twenty physical partitions already in a drive.
Whether it is the right solution for you or someoneelse is a different issue. I was responding to the suggestion that LVM is not safe. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Aaron Kulkis wrote:
But it forces to repartition the entire drive, and I don't feel LVM is as safe. It's not, which is why I don't use it.
All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
Any other suggestions for something that will let me expand my filesystem on my systems that must be kept running 365x24?
None. And that's one of the only ways (the other being lots of Oracle or other database raw partitions) that I would advise using LVM without any reservations.
LVM and EVMS are both actually used quite a lot because they solve a real problem. Similar systems have been in use in e.g. AIX and HPUX for over 10 years.
I know. I've been using them since 1996. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 15:37 -0500, Aaron Kulkis wrote:
Carlos E. R. wrote:
You might consider using LVM as, IIRC, it's not affected by that limit.
But it forces to repartition the entire drive, and I don't feel LVM is as safe.
It's not, which is why I don't use it.
All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
Where is that table stored? Can it be backed up and restored, perhaps? Does it change often? - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHurSktTMYHG2NR9URAr1MAJ4rCOY1TQmYtPrY6oWyos31owY+mgCbB5Em /DEKBS7OHd6nCDHygkRRH0E= =D3Z8 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Monday 2008-02-18 at 15:37 -0500, Aaron Kulkis wrote:
Carlos E. R. wrote:
You might consider using LVM as, IIRC, it's not affected by that limit.
But it forces to repartition the entire drive, and I don't feel LVM is as safe.
It's not, which is why I don't use it.
All LVM partitions rely on the filesystem holding the LVM table to remain uncorrupted. Corruption on *THAT* filesystem can very well destroy your LVM partitioning. :-/
Where is that table stored?
In a file... typically somewhere in /etc or /etc/something_or_other
Can it be backed up and restored, perhaps?
Yes.
Does it change often?
That's totally site-dependant. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 20:37 -0500, Aaron Kulkis wrote:
Where is that table stored?
In a file... typically somewhere in /etc or /etc/something_or_other
But then, that table is inside the lvm. If the table breaks, the LVM breaks, and thus can not be restored, correct? I mean, is there a procedure to restore that file, from a live system and a backup, in order to reconstruct a broken LVM? You see, one of the things that scares me of LVM is that I don't know how to solve it crashes.
Can it be backed up and restored, perhaps?
Yes.
Does it change often?
That's totally site-dependant.
Ie, it changes when the admin changes sizes - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHvAtCtTMYHG2NR9URAiY9AJ9IqksSH5FEtpvkJDSDhrVeB1QKEQCePs00 FtsuhksslWiuIqeP7gLvDRo= =WLE+ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Tuesday 2008-02-19 at 20:37 -0500, Aaron Kulkis wrote:
Where is that table stored?
In a file... typically somewhere in /etc or /etc/something_or_other
But then, that table is inside the lvm. If the table breaks, the LVM breaks, and thus can not be restored, correct? I mean, is there a procedure to restore that file, from a live system and a backup, in order to reconstruct a broken LVM?
You see, one of the things that scares me of LVM is that I don't know how to solve it crashes.
A rigorous backup schedule with a well-tested restore procedure. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
James Knott wrote:
Bob S wrote:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
Has anybody heard anything?? Somebody really working on it? I'll bet not, and we will all be left stuck with this "improvement".
While in some ways I agree with your sentiments, the question I have is what would you do if that was a SCSI drive? Wouldn't that limit still apply regardless?
Yes. However, for the same price, SCSI drives tend to be significantly smaller. Typically if someone is buying 500GB SCSI drives, they're using it for a huge corporate-wide data-base, and using LVM anyways.
You might consider using LVM as, IIRC, it's not affected by that limit.
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Monday 18 February 2008 08:15:22 am James Knott wrote:
Bob S wrote:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
Has anybody heard anything?? Somebody really working on it? I'll bet not, and we will all be left stuck with this "improvement".
While in some ways I agree with your sentiments, the question I have is what would you do if that was a SCSI drive? Wouldn't that limit still apply regardless?
You might consider using LVM as, IIRC, it's not affected by that limit.
Well I have several different OS's in different versions on 3 different drives. I still have 10.2 running and it is on LVM. When I installed 10.3 I was going to use LVM also until I saw that it wanted to combine the 10.2 and the 10.3 volumes. Backed out of that proposal quickly. I want my versions separate and clean. Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 00:02 -0500, Bob S wrote:
You might consider using LVM as, IIRC, it's not affected by that limit.
Well I have several different OS's in different versions on 3 different drives. I still have 10.2 running and it is on LVM. When I installed 10.3 I was going to use LVM also until I saw that it wanted to combine the 10.2 and the 10.3 volumes. Backed out of that proposal quickly. I want my versions separate and clean.
Ouch. That's not nice! If one system filesystem gets corrupted, all will be. So you can not use one to repair the other. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHurUxtTMYHG2NR9URAoRtAJ9mPlqLmULG0l577jhltB32wIRDbQCeJs9h 4hwI+VdJaR3tAWhZbUPFhvQ= =mmhS -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2008/02/17 23:49 (GMT-0500) Bob S apparently typed:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
The kernel devs apparently think lvm and dm are able to provide solutions, that it isn't the job of the kernel to provide backward compatibility for upgrading systems that already have partitions >15. :-(
Has anybody heard anything?? Somebody really working on it?
Rumour has it Device Mapper has the capability, but no pointer to comprehensible docs on how to use dm for the purpose have crossed my path. -- "For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life." John 3:16 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Monday 18 February 2008 09:37:03 am Felix Miata wrote:
On 2008/02/17 23:49 (GMT-0500) Bob S apparently typed:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
The kernel devs apparently think lvm and dm are able to provide solutions, that it isn't the job of the kernel to provide backward compatibility for upgrading systems that already have partitions >15. :-(
Nice of the kernel devs to decide what is best for all of us.
Has anybody heard anything?? Somebody really working on it?
Rumour has it Device Mapper has the capability, but no pointer to comprehensible docs on how to use dm for the purpose have crossed my path. -- So here we sit. Guys like Carlos and I with big drives several OS's and no space to use those big drives in a proper, useful and safe manner. I have 125 GB of free space and two partitions to put 11.0 on. I really don't think anyone cares and they do whatever they want without considering the consequences. It really ticked me off when all of my drives were renumbered.
Please Joe, If you come across a solution please let this list know. Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Bob S wrote:
On Monday 18 February 2008 09:37:03 am Felix Miata wrote:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work! The kernel devs apparently think lvm and dm are able to provide solutions,
On 2008/02/17 23:49 (GMT-0500) Bob S apparently typed: that it isn't the job of the kernel to provide backward compatibility for upgrading systems that already have partitions >15. :-(
Nice of the kernel devs to decide what is best for all of us.
Yeah. Someone needs to dust off the clue-bat.
Has anybody heard anything?? Somebody really working on it? Rumour has it Device Mapper has the capability, but no pointer to comprehensible docs on how to use dm for the purpose have crossed my path. -- So here we sit. Guys like Carlos and I with big drives several OS's and no space to use those big drives in a proper, useful and safe manner. I have 125 GB of free space and two partitions to put 11.0 on. I really don't think anyone cares and they do whatever they want without considering the consequences. It really ticked me off when all of my drives were renumbered.
Please Joe, If you come across a solution please let this list know.
There's a reason I still have 4 actual SCSI hard drives in my system. And I would have 2 more SCSI optical drives, too, if it weren't for a shortage of PCI slots -- i'm not putting those 8-bit single-ended SCSI-2 drives on the same bus as my 16-bit LVD SCSI-3 drives...and having the SCSI-3 drives hampered by adapting to being on the same bus as single-ended drives. In the long term...I don't know what I'm going to do if they don't get this all straightened out. And I'm not sure that I'm all too enthused about these new motherboards. Hell, I'm still annoyed by the industry's decision to reduce the maximum number of card-slots on case and motherboard from 8 to 7. The last incarnation of my desktop machine had all 7 slots used, and I would have used an 8th if only an 8th slot were available. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Feb 19, 2008 12:18 AM, Bob S <911@sanctum.com> wrote:
On Monday 18 February 2008 09:37:03 am Felix Miata wrote:
On 2008/02/17 23:49 (GMT-0500) Bob S apparently typed:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
The kernel devs apparently think lvm and dm are able to provide solutions, that it isn't the job of the kernel to provide backward compatibility for upgrading systems that already have partitions >15. :-(
Nice of the kernel devs to decide what is best for all of us.
At least for now YOU are in control, not the kernel devs. They have taken nothing away and the traditional IDE drivers are being very actively supported. Suse has them in their kernels for 10.3 (11.0 I don't know about). The kernel devs added a parallel development track a copy years ago, so now there are 2 drivers for the vast majority of PATA (IDE) controllers. It is YOUR choice whether you use the traditional set or the new set, nobody else is making decisions for you. So if you don't like the new way, use the old one. (As described in the release notes from day one.) Obviously, SUSE (Novell) had to make a choice of what the default driver set for 10.3 should be. They choose the leading edge set. That is pretty consistent with the opensuse philosophy from what I've seen. I for one am very happy to see them moving forward, and if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries. I really don't understand the gripping about 10.3 in this respect. If suse ever drops the traditional drivers/ide support, that is the time to complain. And those complaints should occur during the alpha/beta process, not 6 months after the product is released. Greg -- Greg Freemyer Litigation Triage Solutions Specialist http://www.linkedin.com/in/gregfreemyer First 99 Days Litigation White Paper - http://www.norcrossgroup.com/forms/whitepapers/99%20Days%20whitepaper.pdf The Norcross Group The Intersection of Evidence & Technology http://www.norcrossgroup.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 09:50 -0500, Greg Freemyer wrote:
At least for now YOU are in control, not the kernel devs.
They have taken nothing away and the traditional IDE drivers are being very actively supported. Suse has them in their kernels for 10.3 (11.0 I don't know about).
The kernel devs added a parallel development track a copy years ago, so now there are 2 drivers for the vast majority of PATA (IDE) controllers. It is YOUR choice whether you use the traditional set or the new set, nobody else is making decisions for you.
So if you don't like the new way, use the old one. (As described in the release notes from day one.)
Obviously, SUSE (Novell) had to make a choice of what the default driver set for 10.3 should be. They choose the leading edge set. That is pretty consistent with the opensuse philosophy from what I've seen. I for one am very happy to see them moving forward, and if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
I really don't understand the gripping about 10.3 in this respect. If suse ever drops the traditional drivers/ide support, that is the time to complain. And those complaints should occur during the alpha/beta process, not 6 months after the product is released.
We are not complaining about 10.3 now. We did complain about 10.3 beta when it was the time: it is on record, if you want to check (bugzillas included). What we complain now is that the written intention is/was to remove the "traditional set" for 11.0 As for the new method being "leading edge"... en fin. Not in one respect. They could have devised a really new device node set and drivers, with neither the limitations of scsi nor ata. They choose compatibility with one set against the other set of device nodes. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuv8KtTMYHG2NR9URAlRGAJ9edNz9MdSOQYZCRsGWFCV3jORbYACdGpjc BjVBXYE6bBQpvd07lcV1k9U= =xD/O -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Feb 19, 2008 11:08 AM, Carlos E. R.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Tuesday 2008-02-19 at 09:50 -0500, Greg Freemyer wrote:
At least for now YOU are in control, not the kernel devs.
They have taken nothing away and the traditional IDE drivers are being very actively supported. Suse has them in their kernels for 10.3 (11.0 I don't know about).
The kernel devs added a parallel development track a copy years ago, so now there are 2 drivers for the vast majority of PATA (IDE) controllers. It is YOUR choice whether you use the traditional set or the new set, nobody else is making decisions for you.
So if you don't like the new way, use the old one. (As described in the release notes from day one.)
Obviously, SUSE (Novell) had to make a choice of what the default driver set for 10.3 should be. They choose the leading edge set. That is pretty consistent with the opensuse philosophy from what I've seen. I for one am very happy to see them moving forward, and if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
I really don't understand the gripping about 10.3 in this respect. If suse ever drops the traditional drivers/ide support, that is the time to complain. And those complaints should occur during the alpha/beta process, not 6 months after the product is released.
We are not complaining about 10.3 now. We did complain about 10.3 beta when it was the time: it is on record, if you want to check (bugzillas included). What we complain now is that the written intention is/was to remove the "traditional set" for 11.0
Don't remember reading that written intention, and per Felix 11.0 still has drivers/ide support. You may be remembering a post I made some time ago guessing that drivers/ide support might be dropped in 11.0 if libata was able to be a full replacement by that time.
As for the new method being "leading edge"... en fin. Not in one respect. They could have devised a really new device node set and drivers, with neither the limitations of scsi nor ata. They choose compatibility with one set against the other set of device nodes.
The new libata drivers leveraged the SCSI infrastructure to get a solid leg up. The have a stated goal to eliminate that reliance and build there own infrastructure. Nobody has actually claimed to be working on that afaik, but it is the stated goal. ((Has been since 2.6.0 I believe, so it has been a very long time coming.) Given that the old drivers/ide is being very actively supported, I suspect Suse will provide us with both options for a while longer. So if it takes another couple years for libata to be removed from the scsi infrastructure and made a full blown subsystem with no scsi limitations, I can live with that. Especially if we have the old drivers/ide option to fall back on. Greg -- Greg Freemyer Litigation Triage Solutions Specialist http://www.linkedin.com/in/gregfreemyer First 99 Days Litigation White Paper - http://www.norcrossgroup.com/forms/whitepapers/99%20Days%20whitepaper.pdf The Norcross Group The Intersection of Evidence & Technology http://www.norcrossgroup.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 12:13 -0500, Greg Freemyer wrote:
On Feb 19, 2008 11:08 AM, Carlos E. R. <> wrote:
...
We are not complaining about 10.3 now. We did complain about 10.3 beta when it was the time: it is on record, if you want to check (bugzillas included). What we complain now is that the written intention is/was to remove the "traditional set" for 11.0
Don't remember reading that written intention, and per Felix 11.0 still has drivers/ide support.
For instance, here: https://bugzilla.novell.com/show_bug.cgi?id=309070#c11 And there were more, but that's the one i could find.
You may be remembering a post I made some time ago guessing that drivers/ide support might be dropped in 11.0 if libata was able to be a full replacement by that time.
Right now, I don't remember.
As for the new method being "leading edge"... en fin. Not in one respect. They could have devised a really new device node set and drivers, with neither the limitations of scsi nor ata. They choose compatibility with one set against the other set of device nodes.
The new libata drivers leveraged the SCSI infrastructure to get a solid leg up. The have a stated goal to eliminate that reliance and build there own infrastructure. Nobody has actually claimed to be working on that afaik, but it is the stated goal. ((Has been since 2.6.0 I believe, so it has been a very long time coming.)
Given that the old drivers/ide is being very actively supported, I suspect Suse will provide us with both options for a while longer. So if it takes another couple years for libata to be removed from the scsi infrastructure and made a full blown subsystem with no scsi limitations, I can live with that. Especially if we have the old drivers/ide option to fall back on.
I hope you are right. Some distros don't support the "classic" driver. However, people using SATA with big disks can't use the classic driver, and others like me can't think of upgrading our hardware to SATA for the time being. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHu5wUtTMYHG2NR9URAoGFAJ9IXbz4uBsIgXgTmDGlsL65/tLZswCcDnOC aTsvL2XHsRhBnJZvyJoPCc0= =ZxLF -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 19 February 2008 09:50:40 am Greg Freemyer wrote:
On Feb 19, 2008 12:18 AM, Bob S <911@sanctum.com> wrote:
On Monday 18 February 2008 09:37:03 am Felix Miata wrote:
On 2008/02/17 23:49 (GMT-0500) Bob S apparently typed:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
The kernel devs apparently think lvm and dm are able to provide solutions, that it isn't the job of the kernel to provide backward compatibility for upgrading systems that already have partitions >15. :-(
Nice of the kernel devs to decide what is best for all of us.
At least for now YOU are in control, not the kernel devs.
They have taken nothing away and the traditional IDE drivers are being very actively supported. Suse has them in their kernels for 10.3 (11.0 I don't know about).
The kernel devs added a parallel development track a copy years ago, so now there are 2 drivers for the vast majority of PATA (IDE) controllers. It is YOUR choice whether you use the traditional set or the new set, nobody else is making decisions for you.
So if you don't like the new way, use the old one. (As described in the release notes from day one.)
Obviously, SUSE (Novell) had to make a choice of what the default driver set for 10.3 should be. They choose the leading edge set. That is pretty consistent with the opensuse philosophy from what I've seen. I for one am very happy to see them moving forward, and if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
Boot flag?? I am totally unaware of that and have no idea what that does. Will that allow me to add partitions to my 125GB of free space? Can you explain more and point me at some docs/how-to's?? Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2008/02/19 23:30 (GMT-0500) Bob S apparently typed:
On Tuesday 19 February 2008 09:50:40 am Greg Freemyer wrote:
if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
Boot flag?? I am totally unaware of that and have no idea what that does. Will that allow me to add partitions to my 125GB of free space? Can you explain more and point me at some docs/how-to's??
http://lists.opensuse.org/opensuse/2008-02/msg02807.html -- "For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life." John 3:16 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 23:39 -0500, Felix Miata wrote:
On 2008/02/19 23:30 (GMT-0500) Bob S apparently typed:
On Tuesday 19 February 2008 09:50:40 am Greg Freemyer wrote:
if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
Boot flag?? I am totally unaware of that and have no idea what that does. Will that allow me to add partitions to my 125GB of free space? Can you explain more and point me at some docs/how-to's??
I think that "hwprobe=-modules.pata" only works during install. Later you remove that option and it remains active, I don't know how. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHvAhRtTMYHG2NR9URAvz6AKCY4ql0BvgHakfoNJmxV3KILFbcqACcCHsa omI64xvkJO5/Gbpj65XwCNs= =lMg1 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Wednesday 20 February 2008 06:00:23 am Carlos E. R. wrote:
The Tuesday 2008-02-19 at 23:39 -0500, Felix Miata wrote:
On 2008/02/19 23:30 (GMT-0500) Bob S apparently typed:
On Tuesday 19 February 2008 09:50:40 am Greg Freemyer wrote:
if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
Boot flag?? I am totally unaware of that and have no idea what that does. Will that allow me to add partitions to my 125GB of free space? Can you explain more and point me at some docs/how-to's??
I think that "hwprobe=-modules.pata" only works during install. Later you remove that option and it remains active, I don't know how.
OK Carlos, Greg, Felix Is that going to work when I go to install 11.0 on the other half of my empty disk? Allow me to make more partitions? I wonder if anyone has tried it? I guess I could go experiment with the partitioner. That frightens me though. Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2008/02/20 23:42 (GMT-0500) Bob S apparently typed:
hwprobe=-modules.pata
Is that going to work when I go to install 11.0 on the other half of my empty disk? Allow me to make more partitions? I wonder if anyone has tried it?
Tried it? I have Factory (11.0) installed on 8 systems. 3 of the 8 *don't* have at least 15 partitions, and it's only on those 3 that Factory is not installed above #15. I have multiple installations of 10.3 above #15 as well.
I guess I could go experiment with the partitioner. That frightens me though.
I don't experiment with partitioners, or use Linux partitioners, or windoz partitioners, or OS/2 partitioners, or DOS partitioners. (I use a cross platform shareware partitioner - DFSee.) My partitioning is normally 100% finished before I ever start any OS installer. The one exception I recall is the one time I used LVM for the target installation, and let the installer do the subdividing on the previously created LVM partition. -- "For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life." John 3:16 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Wed, Feb 20, 2008 at 11:42 PM, Bob S <911@sanctum.com> wrote:
On Wednesday 20 February 2008 06:00:23 am Carlos E. R. wrote:
The Tuesday 2008-02-19 at 23:39 -0500, Felix Miata wrote:
On 2008/02/19 23:30 (GMT-0500) Bob S apparently typed:
On Tuesday 19 February 2008 09:50:40 am Greg Freemyer wrote:
if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
Boot flag?? I am totally unaware of that and have no idea what that does. Will that allow me to add partitions to my 125GB of free space? Can you explain more and point me at some docs/how-to's??
I think that "hwprobe=-modules.pata" only works during install. Later you remove that option and it remains active, I don't know how.
OK Carlos, Greg, Felix
Is that going to work when I go to install 11.0 on the other half of my empty disk? Allow me to make more partitions? I wonder if anyone has tried it?
I guess I could go experiment with the partitioner. That frightens me though.
Bob S
Bob, Before you assume it will work, you do need to verify 10.3 can at least see your drive via the old drivers. ie. with the "hwprobe=-modules.pata" parameter. Greg -- Greg Freemyer Litigation Triage Solutions Specialist http://www.linkedin.com/in/gregfreemyer First 99 Days Litigation White Paper - http://www.norcrossgroup.com/forms/whitepapers/99%20Days%20whitepaper.pdf The Norcross Group The Intersection of Evidence & Technology http://www.norcrossgroup.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tue, Feb 19, 2008 at 11:30 PM, Bob S <911@sanctum.com> wrote:
On Tuesday 19 February 2008 09:50:40 am Greg Freemyer wrote:
On Feb 19, 2008 12:18 AM, Bob S <911@sanctum.com> wrote:
On Monday 18 February 2008 09:37:03 am Felix Miata wrote:
On 2008/02/17 23:49 (GMT-0500) Bob S apparently typed:
Yeah, when they made the switch there were a bunch of us screaming about the partition limitation. We were told, "Don't Worry" that will be fixed. I've got half of a 250 GB disk that I wanted to install SUSE 11 on but only three partitions. That ain't gonna work!
The kernel devs apparently think lvm and dm are able to provide solutions, that it isn't the job of the kernel to provide backward compatibility for upgrading systems that already have partitions >15. :-(
Nice of the kernel devs to decide what is best for all of us.
At least for now YOU are in control, not the kernel devs.
They have taken nothing away and the traditional IDE drivers are being very actively supported. Suse has them in their kernels for 10.3 (11.0 I don't know about).
The kernel devs added a parallel development track a copy years ago, so now there are 2 drivers for the vast majority of PATA (IDE) controllers. It is YOUR choice whether you use the traditional set or the new set, nobody else is making decisions for you.
So if you don't like the new way, use the old one. (As described in the release notes from day one.)
Obviously, SUSE (Novell) had to make a choice of what the default driver set for 10.3 should be. They choose the leading edge set. That is pretty consistent with the opensuse philosophy from what I've seen. I for one am very happy to see them moving forward, and if I need extra partitions, I can add a boot flag, update my grub setup, and fstab entries.
Boot flag?? I am totally unaware of that and have no idea what that does. Will that allow me to add partitions to my 125GB of free space? Can you explain more and point me at some docs/how-to's??
I've never actually used the boot flag, but it is hwprobe=-modules.pata It is described to some extent in the 10.3 release notes. And per Felix, this is still working in 11.0 alphas. Note that this will not work if your controller does not have a traditional drivers/ide driver. Specifically, a lot of sata controllers have pata emulation mode that is supported with the traditional driver set. But if you have the bios set not to emulate you need a real sata driver. Same is true if your controller does not offer an emulation mode. I don't believe you will find any of those in the traditional drivers/ide subsystem. Greg -- Greg Freemyer Litigation Triage Solutions Specialist http://www.linkedin.com/in/gregfreemyer First 99 Days Litigation White Paper - http://www.norcrossgroup.com/forms/whitepapers/99%20Days%20whitepaper.pdf The Norcross Group The Intersection of Evidence & Technology http://www.norcrossgroup.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2008/02/19 00:18 (GMT-0500) Bob S apparently typed:
So here we sit. Guys like Carlos and I with big drives several OS's and no space to use those big drives in a proper, useful and safe manner. I have 125 GB of free space and two partitions to put 11.0 on. I really don't think anyone cares and they do whatever they want without considering the consequences. It really ticked me off when all of my drives were renumbered.
Please Joe, If you come across a solution please let this list know.
The solution/workaround for accessing all your existing partitions as /dev/hdxx in Factory/11.0 is the same as is was in 10.3, as stated in the 10.3 release notes: hwprobe=-modules.pata given as an installation parameter. Note that current development version of Mandriva also includes a functionally equivalent option, while Fedora beginning at v7 has provided no option at all, building all their kernels with no legacy IDE HD drivers, same as Ubuntu starting with v7.10. -- "For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life." John 3:16 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Felix Miata wrote:
Has anybody heard anything?? Somebody really working on it?
Rumour has it Device Mapper has the capability, but no pointer to comprehensible docs on how to use dm for the purpose have crossed my path.
See my couple of postings about 'kpartx'. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2008/02/21 02:12 (GMT-0500) Per Jessen apparently typed:
Felix Miata wrote:
Has anybody heard anything?? Somebody really working on it?
Rumour has it Device Mapper has the capability, but no pointer to comprehensible docs on how to use dm for the purpose have crossed my path.
See my couple of postings about 'kpartx'.
Then put it in the subject line, instead of If you need more than 16 paritions on your SATA drive, this might work do If you need more than 16 paritions on your SATA drive, kpartx might work or something similar, so an archive subject search can turn it up. The million dollar question, is can the SUSE YaST installer use a >15 kpartx/dm partition on SATA as a / target? If it can, there should be no need to avoid upgrading older systems to newer ones that only have SATA, which is what those of us using many partitions have been needing to do. #2, if #1 is yes, is can the YaST installer do it all by itself? -- "For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life." John 3:16 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Felix Miata wrote:
If you need more than 16 paritions on your SATA drive, kpartx might work
or something similar, so an archive subject search can turn it up.
Which is exactly what I did in my third posting after having confirmed that kpartx does indeed work.
The million dollar question, is can the SUSE YaST installer use a >15 kpartx/dm partition on SATA as a / target?
The other important question is whether the openSUSE install system comes with kpartx?
If it can, there should be no need to avoid upgrading older systems to newer ones that only have SATA, which is what those of us using many partitions have been needing to do. #2, if #1 is yes, is can the YaST installer do it all by itself?
It might be a good suggestion for 11.0. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2008-02-17 at 14:29 +0100, Per Jessen wrote:
Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon.
That's a SATA or PATA limit isn't it?
Both, as the idea now is to use the scsi method (for all disk types regardless of what they are really), which has that limitation. In fact, it is even less. One primary, one extended, the rest are logical, and you waste numbers 3 and 4: 13 partitions are left for real use (1,2,5,..15) - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuEMstTMYHG2NR9URAnIGAJ46+BSs4z+Ke0epqbLUA8XS43OgsgCcD5cY +nDz35gjbjvYKC6G6ne/eSU= =vh+d -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sunday 17 February 2008 08:22:34 am Carlos E. R. wrote:
In fact, it is even less. One primary, one extended, the rest are logical, and you waste numbers 3 and 4: 13 partitions are left for real use (1,2,5,..15)
Extended is entry in partition table, so you can use 3 primary + 1 extended. I use 2 + 1 and nobody is complaining. -- Regards, Rajko. See http://en.opensuse.org/Portal -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2008-02-17 at 08:35 -0600, Rajko M. wrote:
On Sunday 17 February 2008 08:22:34 am Carlos E. R. wrote:
In fact, it is even less. One primary, one extended, the rest are logical, and you waste numbers 3 and 4: 13 partitions are left for real use (1,2,5,..15)
Extended is entry in partition table, so you can use 3 primary + 1 extended. I use 2 + 1 and nobody is complaining.
Of course you can, but it is less flexible. Once that #2 is extended type, it's not so simple to change any of them to be #3 and #4. Resizing gets too complex. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuFt4tTMYHG2NR9URAv1zAJ0WxrEINjgt8QJrdIBYpO9RIqfocwCfRQ3+ Pf7T/hetflJxFjdlin8EdDo= =TBO5 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sunday 17 February 2008 10:06:15 am Carlos E. R. wrote:
The Sunday 2008-02-17 at 08:35 -0600, Rajko M. wrote:
On Sunday 17 February 2008 08:22:34 am Carlos E. R. wrote:
In fact, it is even less. One primary, one extended, the rest are logical, and you waste numbers 3 and 4: 13 partitions are left for real use (1,2,5,..15)
Extended is entry in partition table, so you can use 3 primary + 1 extended. I use 2 + 1 and nobody is complaining.
Of course you can, but it is less flexible. Once that #2 is extended type, it's not so simple to change any of them to be #3 and #4. Resizing gets too complex.
:-) Sure. Though, it depends on what is /dev/sda1. Device Boot Start End Blocks Id System /dev/sda1 1 3135 25173855 7 HPFS/NTFS /dev/sda2 6672 19457 102703514 f W95 Ext'd (LBA) /dev/sda3 * 3135 6672 28410952 83 Linux /dev/sda5 6672 6933 2104483+ 82 Linux swap / Solaris /dev/sda6 6934 9366 19543041 83 Linux /dev/sda7 9367 13014 29302528+ 83 Linux /dev/sda8 13015 18237 41945715 83 Linux /dev/sda9 18237 19457 9807619+ 83 Linux Can you see something unusual with /dev/sda3 ? It was created by squeezing windows more than it wwas right after installation. -- Regards, Rajko. See http://en.opensuse.org/Portal -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2008-02-17 at 11:57 -0600, Rajko M. wrote:
Extended is entry in partition table, so you can use 3 primary + 1 extended. I use 2 + 1 and nobody is complaining.
Of course you can, but it is less flexible. Once that #2 is extended type, it's not so simple to change any of them to be #3 and #4. Resizing gets too complex.
:-) Sure. Though, it depends on what is /dev/sda1.
Device Boot Start End Blocks Id System /dev/sda1 1 3135 25173855 7 HPFS/NTFS /dev/sda2 6672 19457 102703514 f W95 Ext'd (LBA) /dev/sda3 * 3135 6672 28410952 83 Linux /dev/sda5 6672 6933 2104483+ 82 Linux swap / Solaris /dev/sda6 6934 9366 19543041 83 Linux /dev/sda7 9367 13014 29302528+ 83 Linux /dev/sda8 13015 18237 41945715 83 Linux /dev/sda9 18237 19457 9807619+ 83 Linux
Can you see something unusual with /dev/sda3 ? It was created by squeezing windows more than it wwas right after installation.
Suppose I have used 1,2,5,6,..16,17. Do you see and easy manner to convert 5 and 6 to 3 and 4, moving the rest down? They are different types. Now, I guess I can consider myself an expert, and perhaps I could get away with it without resorting to Partition Magic, nor a full reinstall. But I wouldn't bet on it, nor on the average Joe doing it. I would certainly make a full backup, and is not easy for a 300 or 500 GB drive. That done, it would probably be easier and safer to fully reinstall on another drive and back, but I'd need to do the investment in spare hardware... - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHurdutTMYHG2NR9URAs3EAJ9iDiLNJ3/8BN7MNGDoluTup82ujwCfSbPf r5c67SNLLpXCqpDC+jIIfxw= =3KOZ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 02/19/2008 07:03 PM, Carlos E. R. wrote:
Suppose I have used 1,2,5,6,..16,17. Do you see and easy manner to convert 5 and 6 to 3 and 4, moving the rest down?
They are different types. Now, I guess I can consider myself an expert, and perhaps I could get away with it without resorting to Partition Magic, nor a full reinstall. But I wouldn't bet on it, nor on the average Joe doing it. Since I am an average Joe, I can tell you it would not be easy. But I have become very impressed with the power and stability of gparted. I recently deleted sda7 on a system (3rd logical), shrunk the primary
If you are saying to make logical partitions primary, though easy is not what I would call it, I believe it could be done with gparted. partition, expanded sda6 to the end of the primary, and expanded a second primary AFTER sda7 in both directions, without a problem. Very impressive.
I would certainly make a full backup, and is not easy for a 300 or 500 GB drive. That done, it would probably be easier and safer to fully reinstall on another drive and back, but I'd need to do the investment in spare hardware...
After my experiences with gparted, I tried the above with no problems. I have been using the gparted live cd to upgrade Windows disks in laptops to bigger drives with great success as well. But this is so far from the subject... -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon.
That's a SATA or PATA limit isn't it?
(IDE) PATA disk drives can have up to 255 partitions, with everything under SCSI emulation now, the SCSI 15-partition limit is applied to everything ( :-/ ). -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, 17 Feb 2008, Aaron Kulkis wrote:-
Per Jessen wrote:
Carlos E. R. wrote:
Yes, but the snag is that the number of partitions available in linux has been lowered to 15. If you do that and have more than one or two operating systems, you will hit the limit soon. That's a SATA or PATA limit isn't it?
(IDE) PATA disk drives can have up to 255 partitions,
IDE drives can hold upto 63 partitions.
with everything under SCSI emulation now, the SCSI 15-partition limit is applied to everything ( :-/ ).
Unless you continue to use the old IDE drivers, or use LVM. The last system I performed a fresh install upon has the following layout: /dev/sda1 /boot /dev/sda2 swap /dev/sda3 extended /dev/sda5 / /dev/sda6 /root /dev/sda7 /home /dev/sda8 /usr /dev/sda9 /var /dev/sda10 /var/log /dev/sda11 /tmp /dev/sda12 LVM /dev/sdb1 LVM Apart from allowing me grow both /opt, and especially /srv, as required, it also lets me span drives for the file system /media/backups, which holds backups of my sources packages. Regards, David Bolt -- Team Acorn: http://www.distributed.net/ OGR-P2 @ ~100Mnodes RC5-72 @ ~15Mkeys SUSE 10.1 32bit | openSUSE 10.2 32bit | openSUSE 10.3 32bit | openSUSE 11.0a1 SUSE 10.1 64bit | openSUSE 10.2 64bit | openSUSE 10.3 64bit RISC OS 3.6 | TOS 4.02 | openSUSE 10.3 PPC | RISC OS 3.11 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 11:52 -0000, David Bolt wrote:
That's a SATA or PATA limit isn't it?
(IDE) PATA disk drives can have up to 255 partitions,
IDE drives can hold upto 63 partitions.
And more. The 63 limit is a software one. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHubfqtTMYHG2NR9URAt3zAJ9/mijxLtg5r9Cl0n3B37dhbs5rTACgizY/ H8g2GYH+d6vTEStDYmz/p2I= =nKhG -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2008-02-16 at 22:04 -0500, Adam Jimerson wrote:
Can the check be ran manually like you can with reiser? That way instead of doing it a boot every 60 days it can be ran while the system is idle.
Sure! Just use "fsck". But the partition can not be mounted at the time. There is a trick, though: touch /forcefsck and the check will happen automatically on next boot (all partitions). - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuB0PtTMYHG2NR9URAh6nAJ49dARl4u8jb++qSjCUbfOB2uOEhgCfbi+4 X0qulRNwuYwQwRJ0i9AEYjc= =iTu8 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat February 16 2008 13:38:11 Carlos E. R. wrote:
It is bad enough that we don't get any decent feedback in the splash screen ( https://bugzilla.novell.com/show_bug.cgi?id=344271 ), which I can imagine is unsettling to many new users, who will have no clue about what is happening. [...] What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
It's reported already, as I pointed out in the link above. Go there and vote for it, as I did.
Perhaps a method could be devised to warn in advance that a disk check is getting near, and then do it at your convenience. Or when the time comes to have the possibility of saying "tomorrow, please".
I like that. Instead of just telling the user "you are out of luck today - this time the boot process will take a long while", you could ask "it is time to check your file system; this process may take a while; do you want to do it now?" We should add this to the enhancement suggestion in https://bugzilla.novell.com/show_bug.cgi?id=344271 . -- Carlos FL "It is not worth an intelligent man's time to be in the majority. By definition, there are already enough people to do that." - G. H. Hardy -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 16 February 2008 11:58:26 pm Carlos F. Lange wrote:
Perhaps a method could be devised to warn in advance that a disk check is getting near, and then do it at your convenience. Or when the time comes to have the possibility of saying "tomorrow, please".
I like that. Instead of just telling the user "you are out of luck today - this time the boot process will take a long while", you could ask "it is time to check your file system; this process may take a while; do you want to do it now?" We should add this to the enhancement suggestion in https://bugzilla.novell.com/show_bug.cgi?id=344271 .
Excellent idea! I added a comment also. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun February 17 2008 10:31:47 Kai Ponte wrote:
We should add this to the enhancement suggestion in https://bugzilla.novell.com/show_bug.cgi?id=344271 .
Excellent idea!
I added a comment also.
Don't forget that you can also vote. I am not sure if it makes much of a difference, but that is what I do if I simply agree with the report. Your point comparing with password renew was good. -- Carlos FL "It is not worth an intelligent man's time to be in the majority. By definition, there are already enough people to do that." - G. H. Hardy -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
<stuff deleted>
Yes, they are needed. Unfortunately.
What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
I have tried, unsucessfully, to disable the splash screen on my own. What do I need to do to disable the splash screen and even boot in verbose mode as you suggest? Thank you ____________________________________________________________________________________ Never miss a thing. Make Yahoo your home page. http://www.yahoo.com/r/hs -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
* Michael Juntunen
I have tried, unsucessfully, to disable the splash screen on my own.
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
edit "/boot/grub/menu.lst" and change "splash=???" to splash=verbose -- Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://counter.li.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Mon February 18 2008 15:02:33 Michael Juntunen wrote:
<stuff deleted>
Yes, they are needed. Unfortunately.
What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
I have tried, unsucessfully, to disable the splash screen on my own.
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
As the message on screen says, you can hit Esc to remove the splash screen and see the messages underneath. If you want to make this change permanent, go to Yast - System - /etc/sysconfig Editor - System - Boot - SPLASH and change the value to "no". -- Carlos FL "It is not worth an intelligent man's time to be in the majority. By definition, there are already enough people to do that." - G. H. Hardy -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 14:02 -0800, Michael Juntunen wrote:
What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
I have tried, unsucessfully, to disable the splash screen on my own.
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
Edit '/boot/grub/menu.lst'. You will find entries similar to this: title MAIN openSUSE 10.3 (default) root (hd0,5) kernel /vmlinuz root=/dev/disk/by-id/ata-ST3160021A_5JS4VV1F-part6 vga=0x317 hwprobe=-modules.pata resume=/dev/hda5 splash=verbose showopts apic initrd /initrd You need "splash=verbose". If you don't like to edit the file directly, you can do the same from Yast, boot manager. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD4DBQFHugT3tTMYHG2NR9URApQxAJiOCXRDjqrF165qR9L2Tl+EE/XEAKCWEaXa 7uaPNtGQ7us2QcU05mFEmA== =7WNq -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 2008/02/18 23:21 (GMT+0100) Carlos E. R. apparently typed:
The Monday 2008-02-18 at 14:02 -0800, Michael Juntunen wrote:
I have tried, unsucessfully, to disable the splash screen on my own.
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
Edit '/boot/grub/menu.lst'. You will find entries similar to this:
title MAIN openSUSE 10.3 (default) ... kernel /vmlinuz root=/dev/sda6 showopts vga=0x317 splash=verbose apic ...
You need "splash=verbose".
splash=0 & splash=off both get rid of the splash screen entirely -- "For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life." John 3:16 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Michael Juntunen wrote:
<stuff deleted>
Yes, they are needed. Unfortunately.
What could be reported as a bug is the no feedback situation with the splash screen. I always disable the splash screen and boot in verbose mode.
I have tried, unsucessfully, to disable the splash screen on my own.
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
If I want to see the verbose stuff, I just hit the Escape key. If I'm not there to hit the escape key, I'm not there to see the verbose messages either...but if I want to later, I can still hit Escape when I come back. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 22:06 -0500, Aaron Kulkis wrote:
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
If I want to see the verbose stuff, I just hit the Escape key.
If I'm not there to hit the escape key, I'm not there to see the verbose messages either...but if I want to later, I can still hit Escape when I come back.
Supposing it did not lock >:-) Are you sure the esc key works while the thing is fsck-ing a large partition? - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHurJntTMYHG2NR9URArp2AJ9rVPsjZtrNTJlrPuzofbtYTg5+/ACbByeG CVYPOFJ9kKLq6MfMDlgukc0= =7OnX -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 19 February 2008 11:41:40 Carlos E. R. wrote:
Are you sure the esc key works while the thing is fsck-ing a large partition?
IRC it will, I thought that is what I did just recently. What I do remember is a while ago someone phoned me because his machine had locked up. It took me a little while to realise what was really happening. This machine was running headless, so no display, I got hold of the only screen available and connected it, still nothing, the screen could not use the settings from menu.lst. Noticing the disk activity I realised that fsck was probably running so went for a coffee. When I got back the machine had finally booted. What the person had been doing was rebooting because he thought it had locked up, so causing fsck to start again. As soon as it was booted I changed the vga=xxxxx to vga=normal now at least any screen put onto the machine displays something and you can see what's going on. Why do we have to ape MS all the time? The system sends out lots of messages, why hide them behind a picture? -- Dave Cotton -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2008-02-19 at 12:03 +0100, Dave Cotton wrote: ...
What the person had been doing was rebooting because he thought it had locked up, so causing fsck to start again.
As soon as it was booted I changed the vga=xxxxx to vga=normal now at least any screen put onto the machine displays something and you can see what's going on.
Why do we have to ape MS all the time? The system sends out lots of messages, why hide them behind a picture?
Yes, that's right. I think people using headless machines should be able to use a small LCD text display for warnings and such. It would save tons of headaches. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHurrgtTMYHG2NR9URApLaAJwKXQm/Y3sPo0PJp2gvEjjSQ+DODgCfTuPD dC1PTB/ZSivX+0Z+egYZnH8= =Uxv0 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Cotton wrote:
On Tuesday 19 February 2008 11:41:40 Carlos E. R. wrote:
Are you sure the esc key works while the thing is fsck-ing a large partition?
IRC it will, I thought that is what I did just recently.
What I do remember is a while ago someone phoned me because his machine had locked up. It took me a little while to realise what was really happening.
This machine was running headless, so no display, I got hold of the only screen available and connected it, still nothing, the screen could not use the settings from menu.lst. Noticing the disk activity I realised that fsck was probably running so went for a coffee. When I got back the machine had finally booted.
What the person had been doing was rebooting because he thought it had locked up, so causing fsck to start again.
As soon as it was booted I changed the vga=xxxxx to vga=normal now at least any screen put onto the machine displays something and you can see what's going on.
Why do we have to ape MS all the time?
The system sends out lots of messages, why hide them behind a
Yes, I really wish that Linux devs would stop trying to imitate the most perverse software producer on the planet. picture? Absolutely. I *HATE* the MS boot process and the fact that everything is hidden, so that you have no idea what's working right and what is not. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Monday 2008-02-18 at 22:06 -0500, Aaron Kulkis wrote:
What do I need to do to disable the splash screen and even boot in verbose mode as you suggest?
If I want to see the verbose stuff, I just hit the Escape key.
If I'm not there to hit the escape key, I'm not there to see the verbose messages either...but if I want to later, I can still hit Escape when I come back.
Supposing it did not lock >:-)
Are you sure the esc key works while the thing is fsck-ing a large partition?
I usually hit it right after the splash screen pops up.... sometimes 10-20 seconds later. I have no clue if I've ever gone from splash screen -> verbose during the middle of an fsck or not. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions?
No. It's a good thing. In the old days, going back to Unix, fsck was performed on every file system on EVERY boot up. If it weren't for periodic forced checks, you would never be fscking your filesystems at all, until some corruption becomes so bad that: 1) you notice it and 2) it's frustratingly difficult, if not impossible to repair without data loss.
I started using ext3 in my large /home partitions, as recommended by openSUSE (still using Reiser for my smaller RAID1 root partition, because of this strange bug: https://bugzilla.novell.com/show_bug.cgi?id=350992 ) and today I passed again the 60 days limit without checking ext3 and I had to wait 10 minutes for the 200GB partition to be checked.
It is bad enough that we don't get any decent feedback in the splash
When the splash screen comes up, hit the Escape (Esc) key, and you'll get all the feedback you want.
screen ( https://bugzilla.novell.com/show_bug.cgi?id=344271 ), which I
It's not a bug.
can imagine is unsettling to many new users, who will have no clue about what is happening.
What I really question is if these lengthy checks are really needed.
If you don't want to know about filesystem corruption until it's past the point of repaiting, then I guess you don't need them.
My Reiser partition was cleared with a simple journal check (isn't that what journaling systems are about?). Add to that the fact that hard disks are becoming bigger and these forced check times can mean one might need to wait 20 minutes or 1/2 hour to get access to the machine again? Obviously this will happen when you have to leave in 10 minutes and you are just about to print out that vital document and your foot hits the reset switch...
What is the use of investing a lot of effort into reducing the boot time by precious seconds to give users a better Linux experience, if you have to wait 10 minutes for fsck every time you boot (since Linux
Fsck is happening EVERY time you boot? Why are you doing improper shut-downs so that each of the filesystems is marked as "dirty"
allows you to run several months without reboot, until the next kernel patch comes by).
It finds and corrects filesystem corruption while the damage is still painlessly repairable.
There, I vented...
How often do you reboot? akulkis@kulkix:~> uptime 4:43pm up 11 days 18:25, 83 users, load average: 4.15, 3.42, 1.75 akulkis@kulkix:~> -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Aaron Kulkis wrote:
In the old days, going back to Unix, fsck was performed on every file system on EVERY boot up.
Same as today then.
What is the use of investing a lot of effort into reducing the boot time by precious seconds to give users a better Linux experience, if you have to wait 10 minutes for fsck every time you boot (since Linux
Fsck is happening EVERY time you boot?
It does by default. See init-scripts boot.rootfsck and boot.localfs. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2008-02-18 at 10:19 +0100, Per Jessen wrote:
Fsck is happening EVERY time you boot?
It does by default. See init-scripts boot.rootfsck and boot.localfs.
Yes, but not the full check; not for ext2/3 at least, which happens only every so many days or mounts. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHuZHStTMYHG2NR9URAnk8AJ9cEZ1XxWyA6TVgjcUTlh8iI7zzEACglZAN Rn/l1FDcfeV7mheaM4T+8Ko= =f6NC -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos F. Lange wrote:
Anyone else getting annoyed by the periodic check forced in ext3 partitions? [...]
If you want to make sure that there's no check at next boot time, use the "fastboot" mechanism. If you want to make sure that there is a file system check at next boot time (e.g. it's a scheduled downtime), use the "forcefsck" mechanism. If you want to disable all automatic checks for ext3 filesystems, use tune2fs (as already mentioned in other emails). However, usually only a replay of the journal takes place at boot time and the filesystem is marked as clean thereafter. Nevertheless, there could still be a filesystem problem somewhere (caused by whatever reason, e.g. bad disk, bad cable, bad memory, kernel bug, filesystem bug etc - these problems will not mark a filesystem as dirty, i.e. usually these problems go undetected at first), that's why in general it's a good idea to run a complete fsck from time to time (this minimizes potential data loss since problems mentioned above should be detected earlier). In principle, this holds for any filesystem, not just ext3. If you disable automatic filesystem checks completely, then it's up to you to initiate regular checks or hazard the consequences of not doing so, respectively. SuSE's default settings might sometimes be a bit inconvenient, but they are conservative and safe. Th. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (27)
-
Aaron Kulkis
-
Adam Jimerson
-
Bob S
-
Carlos E. R.
-
Carlos F. Lange
-
Clayton
-
Dave Cotton
-
David Bolt
-
David C. Rankin
-
Don Raboud
-
Felix Miata
-
Greg Freemyer
-
James Knott
-
Jerry Houston
-
Joe Morris
-
Joe Sloan
-
John Andersen
-
Kai Ponte
-
Kevin Valko
-
Michael Juntunen
-
Patrick Shanahan
-
Per Jessen
-
Philippe Landau
-
Rajko M.
-
The Magic Nose Goblin
-
Thomas Hertweck
-
Tom Patton