Oleg Gawriloff
2009-May-12 11:48 UTC
[Ocfs2-devel] add error check for ocfs2_read_locked_inode() call
After upgrading from 2.6.28.10 to 2.6.29.3 I've saw following new errors in kernel log: May 12 14:46:41 falcon-cl5 May 12 14:46:41 falcon-cl5 (6757,7):ocfs2_read_locked_inode:466 ERROR: status = -22 Only one node is mounted volumes in cluster: /dev/sde on /home/apache/users/D1 type ocfs2 (rw,_netdev,noatime,heartbeat=local) /dev/sdd on /home/apache/users/D2 type ocfs2 (rw,_netdev,noatime,heartbeat=local) /dev/sdc on /home/apache/users/D3 type ocfs2 (rw,_netdev,noatime,heartbeat=local) The error appears multiple times per second and only numbers is different. May 12 14:47:09 falcon-cl5 (6755,6):ocfs2_read_locked_inode:466 ERROR: status = -22 May 12 14:47:09 falcon-cl5 May 12 14:47:09 falcon-cl5 (6755,2):ocfs2_read_locked_inode:466 ERROR: status = -22 May 12 14:47:09 falcon-cl5 May 12 14:47:09 falcon-cl5 (6755,1):ocfs2_read_locked_inode:466 ERROR: status = -22 What does it means, any suggestions?
Sunil Mushran
2009-May-12 23:37 UTC
[Ocfs2-devel] add error check for ocfs2_read_locked_inode() call
Oleg Gawriloff wrote:> After upgrading from 2.6.28.10 to 2.6.29.3 I've saw following new errors > in kernel log: > May 12 14:46:41 falcon-cl5 > May 12 14:46:41 falcon-cl5 (6757,7):ocfs2_read_locked_inode:466 ERROR: > status = -22 > > Only one node is mounted volumes in cluster: > /dev/sde on /home/apache/users/D1 type ocfs2 > (rw,_netdev,noatime,heartbeat=local) > /dev/sdd on /home/apache/users/D2 type ocfs2 > (rw,_netdev,noatime,heartbeat=local) > /dev/sdc on /home/apache/users/D3 type ocfs2 > (rw,_netdev,noatime,heartbeat=local) > The error appears multiple times per second and only numbers is different. > May 12 14:47:09 falcon-cl5 (6755,6):ocfs2_read_locked_inode:466 ERROR: > status = -22 > May 12 14:47:09 falcon-cl5 > May 12 14:47:09 falcon-cl5 (6755,2):ocfs2_read_locked_inode:466 ERROR: > status = -22 > May 12 14:47:09 falcon-cl5 > May 12 14:47:09 falcon-cl5 (6755,1):ocfs2_read_locked_inode:466 ERROR: > status = -22 > > What does it means, any suggestions?I went thru the relevant code in 2.6.29-3. There has to be another message. That is, other than the one listed. http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.29.y.git;a=blob;f=fs/ocfs2/inode.c;h=229e707bc050629c0ad372b14c75194bf4b97f3d;hb=e18346d68f807c62bb8e5de0f809d3a1a2f093de Did you modify the code? Apply some patches?
Oleg Gawriloff
2009-May-13 09:45 UTC
[Ocfs2-devel] add error check for ocfs2_read_locked_inode() call
Sunil Mushran ?????:> I went thru the relevant code in 2.6.29-3. There has to be another > message. That is, other than the one listed. > http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.29.y.git;a=blob;f=fs/ocfs2/inode.c;h=229e707bc050629c0ad372b14c75194bf4b97f3d;hb=e18346d68f807c62bb8e5de0f809d3a1a2f093de > Did you modify the code? Apply some patches?Nope. Standard vanilla-sources 2.6.29.3 on gentoo. No patches/modifies. barzog at falcon-cl3 ~ $ ls -l /usr/src/linux/fs/ocfs2/inode.c -rw-r--r-- 1 root root 37345 ??? 24 01:12 /usr/src/linux/fs/ocfs2/inode.c If needed I can give you access to this server.
Possibly Parallel Threads
- ocfs2 - Kernel panic on many write/read from both
- Node Recovery locks I/O in two-node OCFS2 cluster (DRBD 8.3.8 / Ubuntu 10.10)
- [PATCH 1/1] OCFS2: add error check for ocfs2_read_locked_inode() call
- hist function in a for loop
- [PATCH 1/2] ocfs2: Fix metaecc error messages