GRUB + LVM + XFS?
A message that is not in the archive (sent off-list or not yet received): 1010427293.29260.3.camel@UberGeek
A message that is not in the archive (sent off-list or not yet received): [email protected]
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Quoted 6 lines · Rupa Schomaker
I'm having toubles with the current CVS version of the XFS kernel. I haven't had time to look at it completely, but... it could be:
1) mismatch between userspace and kernel (what LVM is in the XFS kernel? I'm using the debian lvm10 userspace).
Quoted 7 lines
2) bad config
3) ???
A quick "bt" at boot in kdb shows that devfs is in the call path. No, I haven't had time to write everything down. :( Man, I wish there was an easy to dump the kdb info or an OOPs to disk...
Well all I do is for getting kdb output is to use a serial console and another computer - then cut and past then send ;-).
It is very interesting that devfs is referenced - there was a LVM memory corruption problem and a patch for LVM/devfs - it can be found here:
http://marc.theaimsgroup.com/?l=linux-xfs&m=101031443221058&w=2
- From my experience it seems that LVM 1.0.1rc4(ish) is way more robust than the updated LVM 1.0.1 or even CVS LVM 1.0.1 yesterday. However, not everything works as expected with either. You can find a results table for a few tests I have conducted here: (of course YMMV)
http://marc.theaimsgroup.com/?l=linux-lvm&m=100994364215110&w=2
I think you would be very interested in the discussions and activity that has been occuring on the linux-xfs & linux-lvm lists at the moment regarding XFS and LVM. If you are seeing problems I would be very interested in finding out how your problems differ from mine.
- --
Adrian Head
(Public Key available on request.)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE8OkHO8ZJI8OvSkAcRAsHkAJ9OlEWOQMxiXXsJ0OzsllI7icT8pQCeLfM2 95t821X+y25nYdf3WCDRg9k=
=rjIv
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Quoted 6 lines · Rupa Schomaker
I'm having toubles with the current CVS version of the XFS kernel. I haven't had time to look at it completely, but... it could be:
1) mismatch between userspace and kernel (what LVM is in the XFS kernel? I'm using the debian lvm10 userspace).
Quoted 7 lines
2) bad config
3) ???
A quick "bt" at boot in kdb shows that devfs is in the call path. No, I haven't had time to write everything down. :( Man, I wish there was an easy to dump the kdb info or an OOPs to disk...
Well all I do is for getting kdb output is to use a serial console and another computer - then cut and past then send ;-).
It is very interesting that devfs is referenced - there was a LVM memory corruption problem and a patch for LVM/devfs - it can be found here:
http://marc.theaimsgroup.com/?l=linux-xfs&m=101031443221058&w=2
- From my experience it seems that LVM 1.0.1rc4(ish) is way more robust than the updated LVM 1.0.1 or even CVS LVM 1.0.1 yesterday. However, not everything works as expected with either. You can find a results table for a few tests I have conducted here: (of course YMMV)
http://marc.theaimsgroup.com/?l=linux-lvm&m=100994364215110&w=2
I think you would be very interested in the discussions and activity that has been occuring on the linux-xfs & linux-lvm lists at the moment regarding XFS and LVM. If you are seeing problems I would be very interested in finding out how your problems differ from mine.
- --
Adrian Head
(Public Key available on request.)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE8OkHO8ZJI8OvSkAcRAsHkAJ9OlEWOQMxiXXsJ0OzsllI7icT8pQCeLfM2 95t821X+y25nYdf3WCDRg9k=
=rjIv
-----END PGP SIGNATURE-----
Quoted 3 lines · Adrian Head
Adrian Head
(Public Key available on request.)
So, what is it? I'm tired of seeing the message about an unverifiable signature. :-)
Have fun (if at all possible),
Signature · 4 lines
The best we can hope for concerning the people at large is that they be properly armed. -- Alexander Hamilton -- Eric Hopper ([email protected] http://www.omnifarious.org/~hopper) --
Quoted 18 lines · Adrian Head
Quoted 6 lines · Rupa Schomaker
I'm having toubles with the current CVS version of the XFS kernel. I haven't had time to look at it completely, but... it could be:
1) mismatch between userspace and kernel (what LVM is in the XFS kernel? I'm using the debian lvm10 userspace).
Quoted 7 lines
2) bad config
3) ???
A quick "bt" at boot in kdb shows that devfs is in the call path. No, I haven't had time to write everything down. :( Man, I wish there was an easy to dump the kdb info or an OOPs to disk...
Well all I do is for getting kdb output is to use a serial console and another computer - then cut and past then send ;-).
Quoted 6 lines
It is very interesting that devfs is referenced - there was a LVM memory corruption problem and a patch for LVM/devfs - it can be found here:
http://marc.theaimsgroup.com/?l=linux-xfs&m=101031443221058&w=2From my experience it seems that LVM 1.0.1rc4(ish) is way more robust than the updated LVM 1.0.1 or even CVS LVM 1.0.1 yesterday.
Do you have some input which particular changes cause it? Thanks.
Quoted 20 lines
However, not
everything works as expected with either. You can find a results table for a few tests I have conducted here: (of course YMMV)
http://marc.theaimsgroup.com/?l=linux-lvm&m=100994364215110&w=2I think you would be very interested in the discussions and activity that has been occuring on the linux-xfs & linux-lvm lists at the moment regarding XFS and LVM. If you are seeing problems I would be very interested in finding out how your problems differ from mine.
Signature · 9 lines
Adrian Head (Public Key available on request.) _______________________________________________ linux-lvm mailing list [email protected] http://lists.sistina.com/mailman/listinfo/linux-lvm read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
Signature · 15 lines
Regards,
Heinz -- The LVM Guy --
*** Software bugs are stupid.
Nevertheless it needs not so stupid people to solve them ***
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Heinz Mauelshagen Sistina Software Inc.
Senior Consultant/Developer Am Sonnenhang 11
56242 Marienrachdorf
Germany
[email protected] +49 2626 141200
FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-Quoted 6 lines · Heinz J . Mauelshagen
Quoted 2 lines
From my experience it seems that LVM 1.0.1rc4(ish) is way more robust than the updated LVM 1.0.1 or even CVS LVM 1.0.1 yesterday.
Do you have some input which particular changes cause it? Thanks.
The issue was that when I tried to create an XFS snapshot it would Oops and the snapshot creation would not have been completed. The /dev/vg/lv entries would not be there; however, vgdisplay would show the snapshot as being in use but I could not remove the incomplete snapshot.
When I would reboot using a lvm-1.0.1 kernel (2.4.17-xfs+lvm1.0.1) - the machine would hang during startup; however, if I used a lvm-1.0.1rc4(ish) kernel (2.4.16-xfs+lvm1.0.1rc4) it would fix the problem on the fly during bootup and the snapshot would be put in full operation without problems. This was the only way I found to get the machine up & running again.
I don't have any info on which lines of code might have been the cause. I'm still learning how to dig in and dbug. ;-)
Signature · 3 lines
Adrian Head (Public Key available on request.)
Quoted 5 lines · Adrian Head
The issue was that when I tried to create an XFS snapshot it would Oops and the snapshot creation would not have been completed. The
/dev/vg/lventries would not be there; however, vgdisplay would show the snapshot as being in use but I could not remove the incomplete snapshot.
Quoted 8 lines
When I would reboot using a lvm-1.0.1 kernel (2.4.17-xfs+lvm1.0.1) - the machine would hang during startup; however, if I used a lvm-1.0.1rc4(ish) kernel (2.4.16-xfs+lvm1.0.1rc4) it would fix the problem on the fly during bootup and the snapshot would be put in full operation without problems. This was the only way I found to get the machine up & running again.
I don't have any info on which lines of code might have been the cause. I'm still learning how to dig in and dbug. ;-)
Looking at other threads in this list I think the issues are XFS related. I have tried the latest XFS patches from
ftp://oss.sgi.com/projects/xfs/download/patches/2.4.17/
and the snapshots seem OK to me (that's without the VFS lock patch which XFS doesn't seem to need anyway).
patrick
Quoted 9 lines · Patrick Caulfield
Looking at other threads in this list I think the issues are XFS related. I have tried the latest XFS patches from
ftp://oss.sgi.com/projects/xfs/download/patches/2.4.17/
and the snapshots seem OK to me (that's without the VFS lock patch which XFS doesn't seem to need anyway).
patrick
The actual fix was worked out by Eric of SGI by comparing the Linus tree with the LVM code (I hope I have this correct) and it modifies lvm-snap.c - at least appears to be an interaction problem. The weird thing is that this fixes the problem if I run my LVM tests by hand - if I use a script it still fails.
--- lvm-snap.c Wed Jan 9 16:10:01 2002
+++ /localhome/eric/lvm-1.0.1-works/lvm-snap.c Wed Jan 9 16:07:52 2002
http://marc.theaimsgroup.com/?l=linux-xfs&m=101061519711939&w=2
Signature · 3 lines
Adrian Head (Public Key available on request.)
Quoted 7 lines · Patrick Caulfield
Looking at other threads in this list I think the issues are XFS related. I have tried the latest XFS patches from
ftp://oss.sgi.com/projects/xfs/download/patches/2.4.17/
and the snapshots seem OK to me (that's without the VFS lock patch which XFS doesn't seem to need anyway).
The original issue was getting ext3, resierfs & XFS coexisting. I ran tests with almost all possible combinations of 2.4.17-xfs & lvm-1.0.1rc4 / lvm-1.0.1 CVS. It would always be that one filesystem was fine but others would fail etc...
Signature · 3 lines
Adrian Head (Public Key available on request.)