Red Hat Bugzilla – Bug 76942
Bad: NFS TCP client takes 10 minutes for recovery
Last modified: 2007-11-30 17:06:52 EST
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; Q312461)
Description of problem:
Bad: NFS TCP client takes 10 minutes or more to recovery an active socket after
a NFS server failover.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. mount a file system using mount option nfsvers=3, proto=tcp (For
2. mount /test
3. cd /test
4. dd if=/dev/zero of=testdata
5. Failover a NFS server.
Actual Results: It will take 10 to 12 minutes before the nfs client able to
access the mount point.
Expected Results: After the NFS server completely failover, NFS client should
able to access the mount point.
There is a patch in 2.5 kernel fixing this particular problem for NFS client.
RedHat need to port the patch to AS2.1.
*** This bug has been marked as a duplicate of 77465 ***