0

I have a pretty basic Apache two node cluster with flat files on SAN using OCFS2 file system mounted on both nodes.The website is working as expected, but I see a lot of the following messages in both servers' syslogs:

On Apache 1

[248816.701082] (apache2,62749,1):ocfs2_mknod:459 ERROR: status = -31
[248816.702187] (apache2,62749,1):ocfs2_mkdir:613 ERROR: status = -31

On Apache2

[249043.856561] (apache2,46225,2):ocfs2_mknod:459 ERROR: status = -31
[249043.858633] (apache2,46225,2):ocfs2_mkdir:613 ERROR: status = -31

There are around 6000 such lines logged every single day and the error is always about 

ocfs2_mknod:459 and ocfs2_mkdir:613.

Anyone seen this before? I would really like to get rid of these error messages.

Khaled
  • 36,533
  • 8
  • 72
  • 99
Debianuser
  • 421
  • 4
  • 12
  • 29

2 Answers2

0

Apparently the issue was due to bad sectors(not sure how). A few days after I posted this issue, things took a turn for worse. The system complained bad sectors on the partition, then mounted only as read-only. Eventually it stopped mounting the partition completely and we had to create a new partition on SAN and restore data from backup.

The new partition worked for a while, but now one one of the two nodes, the partition is mounted read-only for some reason. I get the following error when I touch a file: 

touch: cannot touch `test_file': No space left on device

On the other node it works fine though. I can read and write as usual.

After spending too much time on this issue we have decided it is not worth the effort and are now looking into alternatives - Maybe glusterfs, GFS2 or even good old NFS share to hold flat files.

Debianuser
  • 421
  • 4
  • 12
  • 29
0

I had the error "No space left on device" but there is was free space on my ocfs2 filesystem with corosync+pacemaker on suse 11 sp2

 "No space left on device"

I Solved the problem with this http://www.novell.com/support/kb/doc.php?id=7008696

c4f4t0r
  • 5,301
  • 3
  • 31
  • 42