Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
I actually have this running at @expert but only had one response... pretty dead over there... let me clarify things a bit... I had two boxes previously running 7.2 with ext2 box-x (broken box in question) and box-lt (laptop) after upgrading to 8.1, I decided to also upgrade to ext3. After recreating mkinird, box-lt worked without issue but box-x had this fsck problem. This is with the stock 8.1 kernel 2.4.8-26. I was able to resolve box-x problem by re-ordering theglobal section at the top of lilo.conf and adding in read-only into this section (it already was on every individual kernel already... don't know why it also needed to be there). This worked fine until I upgraded box-x to 2.4.8-36.1. I have not upgraded box-lt yet because it is running win4lin and they do not have an upgraded kernel for this yet. A new lilo.conf was automatically created during upgrade (which also did not have the read-only in the global section). I tried putting read-only on the new lilo.conf, I tried using the old lilo-conf only changing kernel versions, and I tried mkinitrd... all failed. This is the status so far. status of df -T on box-x: [root@mom root]# df -T FilesystemType 1k-blocks Used Available Use% Mounted on /dev/hdb2 ext3 3016176 2390920472040 84% / /dev/hda2 ext3 23393 3082 19103 14% /boot none tmpfs 63344 0 63344 0% /dev/shm /dev/hda3 ext3 903052746108110980 88% /home /dev/hda1 vfat 1625700 1098676527024 68% /mnt/windows si-mandrake.domain:/usr/lib/corel nfs 9018696 6470592 2273224 75% /usr/lib/corel si-mandrake.domain:/usr/doc/wpohelpfiles nfs 9018696 6470592 2273224 75% /usr/doc/wpohelpfiles si-mandrake.domain:/usr/include/wine-wpo2000 nfs 9018696 6470592 2273224 75% /usr/include/wine-wpo2000 si-mandrake.domain:/usr/share/fonts/wprequired nfs 9018696 6470592 2273224 75% /usr/share/fonts/wprequired si-mandrake.domain:/etc/fonttastic nfs 9018696 6470592 2273224 75% /etc/fonttasticsi-mandrake.domain:/usr/share/fonts/userfonts nfs 9018696 6470592 2273224 75% /usr/share/fonts/userfonts si-mandrake.domain:/var/log/fonttastic nfs 9018696 6470592 2273224 75% /var/log/fonttastic si-mandrake.domain:/home/mom/.wpo2000 nfs 2530240 2243736183688 93% /home/mom/.wpo2000 --- Vincent Danen [EMAIL PROTECTED] wrote: On Thu Nov 29, 2001 at 09:13:04AM -0800, SI Reasoning wrote: (Let's move this over to expert@ where it belongs) That maybe the case...but I upgraded two systems from 7.2 to 8.1 but only had problems with one of them... I was able to resolve it with the old kernel by mkinitrd and placing read-only in the global section at the beginning of lilo.conf. That does not seem to work with the new kernel. Ok... both systems use ext2? But now that I've seen your other message (about problems with -26mdk), I'm now confused on your definition of new kernel. Exactly what kernel are you using on either machine, and what is the output of df -T? potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) I suspect, and this is by no means concrete, that your problem may lie with the ext2-ext3 conversion. I installed an 8.1 system recently using ext3, but it was formatted as ext3. The two disk system should make no difference, and I can't see as a 7.2-8.1 update would be the cause of the problem. When I have a chance, I'll try installing a 8.0 ext2-based system in vmware and then upgrade to 8.1 and converting the ext2 partitions to ext3. But it may take some days before I am able to do this. -- vdanen (at) mandrakesoft.com, OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD ATTACHMENT part 2 application/pgp-signature = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
That maybe the case...but I upgraded two systems from 7.2 to 8.1 but only had problems with one of them... I was able to resolve it with the old kernel by mkinitrd and placing read-only in the global section at the beginning of lilo.conf. That does not seem to work with the new kernel. --- Vincent Danen [EMAIL PROTECTED] wrote: On Mon Nov 26, 2001 at 09:21:18AM -0800, SI Reasoning wrote: potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) I suspect, and this is by no means concrete, that your problem may lie with the ext2-ext3 conversion. I installed an 8.1 system recently using ext3, but it was formatted as ext3. The two disk system should make no difference, and I can't see as a 7.2-8.1 update would be the cause of the problem. When I have a chance, I'll try installing a 8.0 ext2-based system in vmware and then upgrade to 8.1 and converting the ext2 partitions to ext3. But it may take some days before I am able to do this. -- vdanen (at) mandrakesoft.com, OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD Current Linux kernel 2.4.8-34.1mdk uptime: 9 days 8 hours 19 minutes. ATTACHMENT part 2 application/pgp-signature = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
let me amend this abit. I was speaking of the initial 8.1 kernel, not 2.4.8-34.1mdk. I only updated the problem machine to 2.4.8-34.1mdk as I am waiting for win4lin to release the latest kernel before I do the other one. --- SI Reasoning [EMAIL PROTECTED] wrote: That maybe the case...but I upgraded two systems from 7.2 to 8.1 but only had problems with one of them... I was able to resolve it with the old kernel by mkinitrd and placing read-only in the global section at the beginning of lilo.conf. That does not seem to work with the new kernel. --- Vincent Danen [EMAIL PROTECTED] wrote: On Mon Nov 26, 2001 at 09:21:18AM -0800, SI Reasoning wrote: potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) I suspect, and this is by no means concrete, that your problem may lie with the ext2-ext3 conversion. I installed an 8.1 system recently using ext3, but it was formatted as ext3. The two disk system should make no difference, and I can't see as a 7.2-8.1 update would be the cause of the problem. When I have a chance, I'll try installing a 8.0 ext2-based system in vmware and then upgrade to 8.1 and converting the ext2 partitions to ext3. But it may take some days before I am able to do this. -- vdanen (at) mandrakesoft.com, OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD Current Linux kernel 2.4.8-34.1mdk uptime: 9 days 8 hours 19 minutes. ATTACHMENT part 2 application/pgp-signature = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1 = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
On Thu Nov 29, 2001 at 09:13:04AM -0800, SI Reasoning wrote: (Let's move this over to expert@ where it belongs) That maybe the case...but I upgraded two systems from 7.2 to 8.1 but only had problems with one of them... I was able to resolve it with the old kernel by mkinitrd and placing read-only in the global section at the beginning of lilo.conf. That does not seem to work with the new kernel. Ok... both systems use ext2? But now that I've seen your other message (about problems with -26mdk), I'm now confused on your definition of new kernel. Exactly what kernel are you using on either machine, and what is the output of df -T? potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) I suspect, and this is by no means concrete, that your problem may lie with the ext2-ext3 conversion. I installed an 8.1 system recently using ext3, but it was formatted as ext3. The two disk system should make no difference, and I can't see as a 7.2-8.1 update would be the cause of the problem. When I have a chance, I'll try installing a 8.0 ext2-based system in vmware and then upgrade to 8.1 and converting the ext2 partitions to ext3. But it may take some days before I am able to do this. -- vdanen (at) mandrakesoft.com, OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD msg48357/pgp0.pgp Description: PGP signature
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
I had the usb hangup happen at least once on shutdown of the Tyan Trinity AT100 (the system with the fsck problem). --- Randy Welch [EMAIL PROTECTED] wrote: George Mitchell wrote: I know that this is not an answer to your problem, but I am running 8.1 (with 2.4.8-34.1) with all but my home partition formatted ext3. I am constantly crashing this system due to some usb problem jamming up the shutdown. Thus I am continually recovering those partitions from the journals and have never encountered the problem you are describing. I am just wondering if there is something unique about your system that is causing the problem. In any case, I have not seen it (yet). I've seen the USB shutdown problem. This on a ASUS P2B-D. I had the misfortune to have it catch reiserfs in a weak moment where I had to re-install since /etc/fstab got corrupted. Haven't seen usb shutdown problems yet on my new machine. (Tyan Thunder i860... If any kernel hacks want a 860 system to try things out on... ) -randy = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
Juan Quintela [EMAIL PROTECTED] writes: Let me guess: - 7.2 uses as kernel 2.2. - there is a way to generate a kernel in 2.2 2.4 making the fs incompatible with 2.2 (I think that it is the sparse super capability). FWIW I'm using a 2.2 kernel with a partition created with a Mandrake-7.0 and tune2fs -l reports the sparse_super capability: Filesystem features: filetype sparse_super -- Guillaume Cottenceau - http://people.mandrakesoft.com/~gc/
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
guillaume == Guillaume Cottenceau [EMAIL PROTECTED] writes: guillaume Juan Quintela [EMAIL PROTECTED] writes: Let me guess: - 7.2 uses as kernel 2.2. - there is a way to generate a kernel in 2.2 2.4 making the fs incompatible with 2.2 (I think that it is the sparse super capability). guillaume FWIW I'm using a 2.2 kernel with a partition created with a Mandrake-7.0 guillaume and tune2fs -l reports the sparse_super capability: guillaume Filesystem features: filetype sparse_super clearly I need a new crystal ball :( Later, Juan. -- In theory, practice and theory are the same, but in practice they are different -- Larry McVoy
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
George Mitchell wrote: I know that this is not an answer to your problem, but I am running 8.1 (with 2.4.8-34.1) with all but my home partition formatted ext3. I am constantly crashing this system due to some usb problem jamming up the shutdown. Thus I am continually recovering those partitions from the journals and have never encountered the problem you are describing. I am just wondering if there is something unique about your system that is causing the problem. In any case, I have not seen it (yet). I've seen the USB shutdown problem. This on a ASUS P2B-D. I had the misfortune to have it catch reiserfs in a weak moment where I had to re-install since /etc/fstab got corrupted. Haven't seen usb shutdown problems yet on my new machine. (Tyan Thunder i860... If any kernel hacks want a 860 system to try things out on... ) -randy
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) --- George Mitchell [EMAIL PROTECTED] wrote: I know that this is not an answer to your problem, but I am running 8.1 (with 2.4.8-34.1) with all but my home partition formatted ext3. I am constantly crashing this system due to some usb problem jamming up the shutdown. Thus I am continually recovering those partitions from the journals and have never encountered the problem you are describing. I am just wondering if there is something unique about your system that is causing the problem. In any case, I have not seen it (yet). -George Mitchell SI Reasoning wrote: I have 2 hard drives. hda2 is /boot and hdb2 is / both of these run ext3 which were built from former ext2 partitions. When I first did the conversion I had to mkinitrd before I could get / to use ext3. It had worked fine until the latest upgrade to kernel 2.4.8-34.1mdk. Now it kicks out trying to do an fsck because hdb2 is already mounted. Earlier in the boot process it loaded the ext3 module and checked the journal for hdb2, then shortly after loading devfs, it again tries to fsck hdb2 and halts the process with the error message stating that hdb2 is already mounted. I have tried creating a new mkinitrd with the 2.4.8-34.1mdk kernel but I still get this message. Any ideas or is this a known bug? I have researched the lists and could not find this mentioned yet. = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1 = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
SI Reasoning wrote: potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) I am using k6-2 450 so you can eliminate that factor, but I find the fact that you upgraded from ext2 interesting. I wonder if the result is the same as a clean ext3 install or if there is some sort of kudging involved? / and /boot on separate drives could be confusing something I suppose and the fact that you went the upgrade route and not the clean install route, well upgrades are notorious for wierd problems but I think a connection with disk audit problems would be unlikely, assuming your initscripts is current for 8.1. Hopefully someone on the list can spot some pattern here and come up with a solution. -George --- George Mitchell [EMAIL PROTECTED] wrote: I know that this is not an answer to your problem, but I am running 8.1 (with 2.4.8-34.1) with all but my home partition formatted ext3. I am constantly crashing this system due to some usb problem jamming up the shutdown. Thus I am continually recovering those partitions from the journals and have never encountered the problem you are describing. I am just wondering if there is something unique about your system that is causing the problem. In any case, I have not seen it (yet). -George Mitchell SI Reasoning wrote: I have 2 hard drives. hda2 is /boot and hdb2 is / both of these run ext3 which were built from former ext2 partitions. When I first did the conversion I had to mkinitrd before I could get / to use ext3. It had worked fine until the latest upgrade to kernel 2.4.8-34.1mdk. Now it kicks out trying to do an fsck because hdb2 is already mounted. Earlier in the boot process it loaded the ext3 module and checked the journal for hdb2, then shortly after loading devfs, it again tries to fsck hdb2 and halts the process with the error message stating that hdb2 is already mounted. I have tried creating a new mkinitrd with the 2.4.8-34.1mdk kernel but I still get this message. Any ideas or is this a known bug? I have researched the lists and could not find this mentioned yet. = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1 = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
george == George Mitchell [EMAIL PROTECTED] writes: george SI Reasoning wrote: potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) george I am using k6-2 450 so you can eliminate that factor, but I find the george fact that you upgraded from ext2 interesting. I wonder if the result george is the same as a clean ext3 install or if there is some sort of george kudging involved? / and /boot on separate drives could be confusing george something I suppose and the fact that you went the upgrade route and george not the clean install route, well upgrades are notorious for wierd george problems but I think a connection with disk audit problems would be george unlikely, assuming your initscripts is current for 8.1. Hopefully george someone on the list can spot some pattern here and come up with a george solution. Hi Let me guess: - 7.2 uses as kernel 2.2. - there is a way to generate a kernel in 2.2 2.4 making the fs incompatible with 2.2 (I think that it is the sparse super capability). could you check/sent me the output of: tune2fs -l your partition please? Later, Juan. -- In theory, practice and theory are the same, but in practice they are different -- Larry McVoy
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
[root@mom root]# tune2fs -l /dev/hdb2 tune2fs 1.24a (02-Sep-2001)'mom,10.0.0.77' (RSA1) to the list of known hosts. Filesystem volume name: none Last mounted on: not available Filesystem UUID: ca78be02-e630-11d4-938c-ff1614a77075 Filesystem magic number: 0xEF53 Filesystem revision #:1 (dynamic) Filesystem features: has_journal filetype needs_recovery sparse_super Filesystem state: clean Errors behavior: Continue Filesystem OS type: Linux Inode count: 383232 Block count: 766080 Reserved block count: 38304 Free blocks: 157039 Free inodes: 239356 First block: 0 Block size: 4096 Fragment size:4096 Blocks per group: 32768 Fragments per group: 32768 Inodes per group: 15968 Inode blocks per group: 499 Last mount time: Sun Nov 25 23:17:43 2001 Last write time: Sun Nov 25 23:17:43 2001 Mount count: 18 Maximum mount count: 20 Last checked: Sun Nov 25 01:32:20 2001 Check interval: 15552000 (6 months) Next check after: Fri May 24 02:32:20 2002 Reserved blocks uid: 0 (user root) Reserved blocks gid: 0 (group root) First inode: 11 Inode size: 128 Journal UUID: none Journal inode:8 Journal device: 0x First orphan inode: 368873 --- Juan Quintela [EMAIL PROTECTED] wrote: george == George Mitchell [EMAIL PROTECTED] writes: george SI Reasoning wrote: potential uniqueness: - 2 disk system with /boot on hda2 and / on hdb2 - upgraded mdk7.2 to 8.1 - migrated from ext2 to ext3 (previous kernel) - k6-2 400 cpu (in case it is an amd chipset issue) george I am using k6-2 450 so you can eliminate that factor, but I find the george fact that you upgraded from ext2 interesting. I wonder if the result george is the same as a clean ext3 install or if there is some sort of george kudging involved? / and /boot on separate drives could be confusing george something I suppose and the fact that you went the upgrade route and george not the clean install route, well upgrades are notorious for wierd george problems but I think a connection with disk audit problems would be george unlikely, assuming your initscripts is current for 8.1. Hopefully george someone on the list can spot some pattern here and come up with a george solution. Hi Let me guess: - 7.2 uses as kernel 2.2. - there is a way to generate a kernel in 2.2 2.4 making the fs incompatible with 2.2 (I think that it is the sparse super capability). could you check/sent me the output of: tune2fs -l your partition please? Later, Juan. -- In theory, practice and theory are the same, but in practice they are different -- Larry McVoy = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
Re: [Cooker] (8.1) ext3 fsck problem with latest kernel
I know that this is not an answer to your problem, but I am running 8.1 (with 2.4.8-34.1) with all but my home partition formatted ext3. I am constantly crashing this system due to some usb problem jamming up the shutdown. Thus I am continually recovering those partitions from the journals and have never encountered the problem you are describing. I am just wondering if there is something unique about your system that is causing the problem. In any case, I have not seen it (yet). -George Mitchell SI Reasoning wrote: I have 2 hard drives. hda2 is /boot and hdb2 is / both of these run ext3 which were built from former ext2 partitions. When I first did the conversion I had to mkinitrd before I could get / to use ext3. It had worked fine until the latest upgrade to kernel 2.4.8-34.1mdk. Now it kicks out trying to do an fsck because hdb2 is already mounted. Earlier in the boot process it loaded the ext3 module and checked the journal for hdb2, then shortly after loading devfs, it again tries to fsck hdb2 and halts the process with the error message stating that hdb2 is already mounted. I have tried creating a new mkinitrd with the 2.4.8-34.1mdk kernel but I still get this message. Any ideas or is this a known bug? I have researched the lists and could not find this mentioned yet. = SI Reasoning [EMAIL PROTECTED] They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. Benjamin Franklin __ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1
