[Ocfs2-users] OCFS2 problems, bug

Sunil Mushran sunil.mushran at oracle.com
Mon Nov 16 10:26:11 PST 2009


http://oss.oracle.com/bugzilla/show_bug.cgi?id=1162

The fix was included in 2.6.31. Ask your distro to apply this patch.
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=a1b08e75dff3dc18a88444803753e667bb1d126e

As far as the performance goes, have you done any baseline tests
on the shared device.

dd if=/dev/sdX of=/dev/null bs=4K count=10000 flags=direct

Do on each node and get some stable figures. Then run on all nodes
concurrently. See if there is any drop off and how much.

The other thing to see would be the interconnect. Ensure you have
a gige private interconnect.

Andy Brown wrote:
> Hi all,
>
> Did some searching and couldn't really find the answer here, but I'm constantly getting this shown, and performance on OCFS2 is shocking on the hosts I've been trying it with.
>
> Linux version 2.6.30-2-686 (Debian 2.6.30-8) (waldi at debian.org) (gcc version 4.3.4 (Debian 4.3.4-3) ) #1 SMP Sat Sep 26 01:16:22 UTC 2009
>
>
>
>
> [  618.282121] kernel BUG at /build/buildd-linux-2.6_2.6.30-8-i386-Tm0Lt9/linux-2.6-2.6.30/debian/build/source_i386_none/fs/ocfs2/dlmglue.c:2470!
> [  618.282207] invalid opcode: 0000 [#1] SMP
> [  618.282327] last sysfs file: /sys/module/nfsd/initstate
> [  618.282384] Modules linked in: ocfs2 jbd2 quota_tree ocfs2_dlmfs ocfs2_stack_o2cb ocfs2_dlm ocfs2_nodemanager ocfs2_stackglue configfs aoe nfsd exportfs nfs lockd fscache nfs_acl auth_rpcgss sunrpc loop snd_pcsp snd_pcm shpchp snd_timer pci_hotplug i3000_edac ipmi_si snd ipmi_msghandler soundcore edac_core hpilo joydev rng_core psmouse snd_page_alloc evdev serio_raw container processor button ext3 jbd mbcache dm_mod ide_cd_mod cdrom sd_mod crc_t10dif ide_pci_generic usbhid hid ata_generic uhci_hcd ahci libata piix scsi_mod tg3 libphy ehci_hcd ide_core usbcore thermal fan thermal_sys [last unloaded: configfs]
> [  618.284848]
> [  618.284900] Pid: 1919, comm: nfsd Not tainted (2.6.30-2-686 #1) ProLiant DL320 G5
> [  618.284976] EIP: 0060:[<f8c5070d>] EFLAGS: 00010246 CPU: 0
> [  618.285062] EIP is at ocfs2_dentry_lock+0x25/0xeb [ocfs2]
> [  618.285120] EAX: f641fa00 EBX: f6f9a8d0 ECX: f8c45648 EDX: 00000000
> [  618.285179] ESI: 2a7f074e EDI: f6495000 EBP: 00000005 ESP: f5e49e84
> [  618.285238]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
> [  618.285297] Process nfsd (pid: 1919, ti=f5e48000 task=f68f0280 task.ti=f5e48000)
> [  618.285371] Stack:
> [  618.285422]  00000000 f6f9a8d0 2a7f074e 00000000 f6f9a8d0 f8c68ab0 00000000 2a7f074e
> [  618.285714]  f8c6b2cd f6f9de00 f6495000 f6f9e940 f6f9e5c0 00000000 00000000 00000000
> [  618.286000]  00000000 00000000 00000000 00000000 00000000 f6f98898 f5861060 00000000
> [  618.286000] Call Trace:
> [  618.286000]  [<f8c68ab0>] ? ocfs2_remote_dentry_delete+0xe/0x95 [ocfs2]
> [  618.286000]  [<f8c6b2cd>] ? ocfs2_unlink+0x403/0xa28 [ocfs2]
> [  618.286000]  [<c0199afb>] ? shrink_dcache_parent+0x33/0xe8
> [  618.286000]  [<c0195133>] ? vfs_rmdir+0x66/0x99
> [  618.286000]  [<f884e6ec>] ? nfsd_unlink+0xf4/0x13e [nfsd]
> [  618.286000]  [<f8854a2f>] ? nfsd3_proc_rmdir+0x62/0x6d [nfsd]
> [  618.286000]  [<f884b24d>] ? nfsd_dispatch+0xdc/0x1a8 [nfsd]
> [  618.286000]  [<f86f87b7>] ? svc_process+0x38c/0x59f [sunrpc]
> [  618.286000]  [<f884b6d7>] ? nfsd+0xde/0x11d [nfsd]
> [  618.286000]  [<f884b5f9>] ? nfsd+0x0/0x11d [nfsd]
> [  618.286000]  [<c01368dc>] ? kthread+0x42/0x67
> [  618.286000]  [<c013689a>] ? kthread+0x0/0x67
> [  618.286000]  [<c0103a97>] ? kernel_thread_helper+0x7/0x10
> [  618.286000] Code: ff e9 19 fe ff ff 55 57 56 53 83 ec 04 83 fa 01 8b 50 58 19 ed 83 e5 fe 83 c5 05 89 14 24 8b 40 54 85 d2 8b b8 98 01 00 00 75 04 <0f> 0b eb fe 8d 9f 9c 00 00 00 89 d8 e8 54 e2 6c c7 8b 87 a4 00
> [  618.286000] EIP: [<f8c5070d>] ocfs2_dentry_lock+0x25/0xeb [ocfs2] SS:ESP 0068:f5e49e84
> [  618.290341] ---[ end trace 7942be0959cd685f ]---
>   



More information about the Ocfs2-users mailing list