[Ocfs2-users] ocfs2_get_suballoc_slot_bit and ocfs2_test_inode_bit on NFS shared ocfs2 filesystem

[David Brezynski]

Hello - I'm hoping someone can help with this.

I'm seeing the errors below in a 3 node cluster with 4 ocfs2-mounted filesystems.  The errors started about a week ago.

System specifics (all 3 cluster nodes):

RHEL 5.6 (kernel 2.6.18-238.el5)
OCFS versions:
rpm -qa| grep ocfs
ocfs2console-1.4.4-1.el5
ocfs2-2.6.18-128.7.1.el5-1.4.7-1.el5
ocfs2-tools-1.4.4-1.el5
ocfs2-2.6.18-238.el5-1.4.7-1.el5

Three node cluster, NFS sharing OCFS2 filesystems from node 1 of cluster to 3 servers not part of the cluster.  I only see the errors on the node which is NFS-exporting the ocfs2 filesystems to other servers.  No
errors on the 2nd/3rd cluster nodes.  

I see a few references to similar errors (no resolution), for example:  http://comments.gmane.org/gmane.comp.file-systems.ocfs2.user/5590
And this bugzilla report references the 2 errors: https://oss.oracle.com/bugzilla/show_bug.cgi?id=1235

Errors from /var/log/messages (I see about 40 an hour) - all messages refer to the same inode (258058):

Jul  3 13:18:05 XXXXX kernel: (nfsd,9554,2):ocfs2_get_suballoc_slot_bit:2103 ERROR: invalid inode 258058 requested
Jul  3 13:18:05 XXXXX kernel: (nfsd,9554,2):ocfs2_test_inode_bit:2207 ERROR: get alloc slot and bit failed -22
Jul  3 13:18:10 XXXXX kernel: (nfsd,9585,2):ocfs2_get_suballoc_slot_bit:2103 ERROR: invalid inode 258058 requested
Jul  3 13:18:10 XXXXX kernel: (nfsd,9585,2):ocfs2_test_inode_bit:2207 ERROR: get alloc slot and bit failed -22
Jul  3 13:18:35 XXXXX kernel: (nfsd,9554,2):ocfs2_get_suballoc_slot_bit:2103 ERROR: invalid inode 258058 requested
Jul  3 13:18:35 XXXXX kernel: (nfsd,9554,2):ocfs2_test_inode_bit:2207 ERROR: get alloc slot and bit failed -22
Jul  3 13:20:10 XXXXX kernel: (nfsd,9585,2):ocfs2_get_suballoc_slot_bit:2103 ERROR: invalid inode 258058 requested
Jul  3 13:20:10 XXXXX kernel: (nfsd,9585,2):ocfs2_test_inode_bit:2207 ERROR: get alloc slot and bit failed -22
Jul  3 13:22:10 XXXXX kernel: (nfsd,9595,0):ocfs2_get_suballoc_slot_bit:2103 ERROR: invalid inode 258058 requested
Jul  3 13:22:10 XXXXX kernel: (nfsd,9595,0):ocfs2_test_inode_bit:2207 ERROR: get alloc slot and bit failed -22

I'm planning on getting downtime to force a fsck.ocfs2 this weekend.  Is this the best course of action?  The filesystem in question is about 800 GB - how long should I expect fsck to run (probably depends on
number of errors found!).

Thanks very much

David Brezynski
TSO UNIX Team
UW Medicine
Seattle WA
Email:       brezy at uw.edu
Desk phone:  (206)616-6119