Justin Piszcz wrote:
> On Sat, 17 Oct 2009, Justin Piszcz wrote:
>> I have a system I recently upgraded from 2.6.30.x and after
>> approximately 24-48 hours--sometimes longer, the system cannot write
>> any more files to disk (luckily though I can still write to /dev/shm)
>> -- to which I have
>> saved the sysrq-t and sysrq-w output:
Unfortunately it looks like a lot of the sysrq-t, at least, was lost.
The sysrq-w trace has the "show blocked state" start a ways down the file,
for anyone playing along at home ;)
Other things you might try are a sysrq-m to get memory state...
>> $ cat /proc/mdstat Personalities : [raid1] [raid6] [raid5] [raid4] md1
>> : active raid1 sdb2 sda2
>> 136448 blocks [2/2] [UU]
>> md2 : active raid1 sdb3 sda3
>> 129596288 blocks [2/2] [UU]
>> md3 : active raid5 sdj1 sdi1 sdh1 sdf1 sdg1 sde1
>> sdd1 sdc1
>> 5128001536 blocks level 5, 1024k chunk, algorithm 2 [8/8] [UUUUUUUU]
>> md0 : active raid1 sdb1 sda1
>> 16787776 blocks [2/2] [UU]
>> $ mount
>> /dev/md2 on / type xfs (rw,noatime,nobarrier,logbufs=8,logbsize=262144)
>> tmpfs on /lib/init/rw type tmpfs (rw,nosuid,mode=0755)
>> proc on /proc type proc (rw,noexec,nosuid,nodev)
>> sysfs on /sys type sysfs (rw,noexec,nosuid,nodev)
>> udev on /dev type tmpfs (rw,mode=0755)
>> tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)
>> devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=620)
>> /dev/md1 on /boot type ext3 (rw,noatime)
>> /dev/md3 on /r/1 type xfs
>> rpc_pipefs on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw)
>> nfsd on /proc/fs/nfsd type nfsd (rw)
Do you get the same behavior if you don't add the log options at mount time?
Kind of grasping at straws here for now ...
>> Distribution: Debian Testing
>> Arch: x86_64
>> The problem occurs with 2.6.31 and I upgraded to 126.96.36.199 and the problem
> In addition to using netconsole, which kernel options should be enabled
> to better diagnose this issue?
> Should I enable these to help track down this bug?
> [ ] XFS Debugging support (EXPERIMENTAL)
> [ ] Compile the kernel with frame pointers
The former probably won't hurt; the latter might gibe us better backtraces.
> Are there any other options that will help determine the root cause of this
> bug that are recommended?
Not that I can think of off hand ...