NFSv4, often client crash one at three times per day and I see this error in log "rpciod waiting on sync task!", It's not a really crash, but it's similar, because the user can't do nothing and other users can't login to the station, but root can if Ctrl C is pushed at login. lsof doesn't work and tcpdump show nothing bad. I see a similar problem on bugzilla and Steve Dickson said that it will be fixed in the upcoming release of RHEL4. But I can't go to the bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=194793 Version-Release number of selected component (if applicable): We are using NFSv4 on RedHat 4 Enterprise kernel 2.6.9-42 All clients and servers run on kernel 2.6.9-42, but servers and clusters are 64- bit How reproducible: I can reproduce the bug with the software subversion on a CVS folder. Additional info: Is it a problem with socket ? Is it a problem with open lock file on NFS ?
I'm seeing the same thing. kernel 2.6.9-42.0.8. NFSv4 server and clients are 64-bit. And I can't go to that bug, either. Sure wish there were at least a workaround!
Oh yeah, the rpciod process on the client is in state "D". Is there a known bug where this process gets deadlocked?
I believe I have a patch going into 4.6 that should fix this problem. If you have a non-critical place to do so, please test the kernels on my people page and let me know if they resolve the issue: http://people.redhat.com/jlayton
*** This bug has been marked as a duplicate of 228292 ***