Comment 40 for bug 932687

Revision history for this message
gergnz (gergnz) wrote :

FWIW, just did an upgrade on some systems.

All preices machines, all were running -39

Upgraded to -48, this bug appeared.

Rolled back the clients only to -39, bug disappeared.

Site is a semi heavy web service with 1 NFS server, and 2 NFS clients running nginx/php.

Issue was seen almost immediately upon nginx/php starting up, lots of threads got stuck, load went up due to high I/O wait.

fstab:
10.1.2.3:/www /data/www nfs defaults 0 0
exports:
/data/shared 10.1.2.0/24(rw,fsid=0,insecure,no_subtree_check,async,no_root_squash)
/data/shared/www 10.1.2.0/24(rw,nohide,insecure,no_subtree_check,async,no_root_squash)