Re: [opensuse-factory] Re: XFS crash in Tumbleweed 20160302 / kernel 4.4.3-1.1
2016-03-08 9:55 GMT+01:00 René Krell <renda.krell@gmail.com>:
.. I cannot exclude a hardware problem, but it has been the only "user-visible" problem to this time with this disk. I use it already several months. Another fact is there was a bunch of XFS changes during the transition from 4.4.2 to 4.4.3, just have a look to https://cdn.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.3. Isn't it possible there has been introduced some race or something similar? The crash occurred while doing extensive disk operations (Maven compilation, launching a Java IDE and similar etc. at the same time).
Look especially at this one: commit 7530e6fdd9f207a6ebcf669490656def4f7cf73e Author: Dave Chinner <dchinner@redhat.com> Date: Tue Jan 12 07:03:44 2016 +1100 xfs: inode recovery readahead can race with inode buffer creation commit b79f4a1c68bb99152d0785ee4ea3ab4396cdacc6 upstream. with seems to have something in common with the trace:
2016-03-04T09:34:42.592321+01:00 rkrell kernel: [ 3442.618298] XFS (sdb2): xfs_do_force_shutdown(0x8) called from line 991 of file ../fs/xfs/xfs_trans.c. Return address = 0xffffffffa12efdef
Just have a look at the commit: https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/... The committer added: xfs_buf_ioerror(bp, -EIO); and in my trace there is: XFS (sdb2): xfs_log_force: error -5 returned. http://www-numi.fnal.gov/offline_software/srt_public_context/WebDocs/Errors/...: #define EIO 5 /* I/O error */ Doesn't this indicate the appropriate commit to be strongly related to this crash? 2016-03-08 9:59 GMT+01:00 René Krell <renda.krell@gmail.com>:
2016-03-08 9:55 GMT+01:00 René Krell <renda.krell@gmail.com>:
.. I cannot exclude a hardware problem, but it has been the only "user-visible" problem to this time with this disk. I use it already several months. Another fact is there was a bunch of XFS changes during the transition from 4.4.2 to 4.4.3, just have a look to https://cdn.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.3. Isn't it possible there has been introduced some race or something similar? The crash occurred while doing extensive disk operations (Maven compilation, launching a Java IDE and similar etc. at the same time).
Look especially at this one:
commit 7530e6fdd9f207a6ebcf669490656def4f7cf73e Author: Dave Chinner <dchinner@redhat.com> Date: Tue Jan 12 07:03:44 2016 +1100
xfs: inode recovery readahead can race with inode buffer creation
commit b79f4a1c68bb99152d0785ee4ea3ab4396cdacc6 upstream.
with seems to have something in common with the trace:
2016-03-04T09:34:42.592321+01:00 rkrell kernel: [ 3442.618298] XFS (sdb2): xfs_do_force_shutdown(0x8) called from line 991 of file ../fs/xfs/xfs_trans.c. Return address = 0xffffffffa12efdef -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello René, On 08.03.2016 12:44, René Krell wrote:
2016-03-08 9:55 GMT+01:00 René Krell <renda.krell@gmail.com>:
.. I cannot exclude a hardware problem, but it has been the only "user-visible" problem to this time with this disk. I use it already several months.
Ok, it's just that these low level error messages look suspicious, and if the disk below the filesystem actually fails, there is not too much the FS can do :-) I had seen those "lost async page write" previously with bad SATA cable/controller/disk/powersupply combinations and with broken USB sticks, hence my suggestion to look there first.
Another fact is there was a bunch of XFS changes during the transition from 4.4.2 to 4.4.3, just have a look to https://cdn.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.4.3. Isn't it possible there has been introduced some race or something similar? The crash occurred while doing extensive disk operations (Maven compilation, launching a Java IDE and similar etc. at the same time).
Look especially at this one:
commit 7530e6fdd9f207a6ebcf669490656def4f7cf73e Author: Dave Chinner <dchinner@redhat.com> Date: Tue Jan 12 07:03:44 2016 +1100
xfs: inode recovery readahead can race with inode buffer creation
commit b79f4a1c68bb99152d0785ee4ea3ab4396cdacc6 upstream.
with seems to have something in common with the trace:
2016-03-04T09:34:42.592321+01:00 rkrell kernel: [ 3442.618298] XFS (sdb2): xfs_do_force_shutdown(0x8) called from line 991 of file ../fs/xfs/xfs_trans.c. Return address = 0xffffffffa12efdef
Just have a look at the commit: https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/...
The committer added: xfs_buf_ioerror(bp, -EIO);
and in my trace there is: XFS (sdb2): xfs_log_force: error -5 returned.
http://www-numi.fnal.gov/offline_software/srt_public_context/WebDocs/Errors/...: #define EIO 5 /* I/O error */
Doesn't this indicate the appropriate commit to be strongly related to this crash?
This certainly might be related. It might be helpful to not only file a bug at bugzilla.kernel.org, but also write to LKML / Dave Chinner, because there is a wide variance in bugzilla usage / attention among kernel developers (some use it, some don't look at it at all). Good luck :-) Stefan -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (2)
-
René Krell -
Stefan Seyfried