Skip to contentSkip to conversation

Expected NFS behavior?

DISCUSSION Bill Johnstone 2 participants 2 messages Feb 17, 2006 – Feb 18, 2006

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hello.

I've got a network wherein the machines are all running kernel 2.6.12, and NFSv3 is exclusively being used. The exports / mounts are all 'sync'. I'm seeing some behavior that doesn't seem to be correct to me, and I want to see if this is expected or if I'm hitting a bug or perhaps hardware failure/flakiness along the way.

The scenario:

A user has an existing set of files on the server, on an XFS volume.

If he is on the server (i.e. logged in via ssh) and runs a cp command to overwrite existing file "a" with file "b", then file "a" appears to be correctly overwritten as verified from the command-line on the server. However, when any of the client machines opens file "a" after this, they see the previous contents. It seems like an incoherent cache issue.

If the user instead runs the cp command on one of the client machines, then all the client machines now see the updated file.

Does this sound like it should be expected behavior, i.e. if the file changes "underneath" NFS (direct to the volume via command-line) will NFS have stale information, or should it explicitly handle this case?

Thanks.

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around=20
http://mail.yahoo.com=20

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi= les
for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D103432&bid=3D230486&dat=3D=
121642
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs

#1 lore raw
Quoted 19 lines · Bill Johnstone

Hello.

I've got a network wherein the machines are all running kernel 2.6.12, and NFSv3 is exclusively being used. The exports / mounts are all 'sync'. I'm seeing some behavior that doesn't seem to be correct to me, and I want to see if this is expected or if I'm hitting a bug or perhaps hardware failure/flakiness along the way.

The scenario:

A user has an existing set of files on the server, on an XFS volume.

If he is on the server (i.e. logged in via ssh) and runs a cp command to overwrite existing file "a" with file "b", then file "a" appears to be correctly overwritten as verified from the command-line on the server. However, when any of the client machines opens file "a" after this, they see the previous contents. It seems like an incoherent cache issue.

Quoted 6 lines

If the user instead runs the cp command on one of the client machines, then all the client machines now see the updated file.

Does this sound like it should be expected behavior, i.e. if the file changes "underneath" NFS (direct to the volume via command-line) will NFS have stale information, or should it explicitly handle this case?

Try a more recent client. There were a number of fixes that went into 2.6.14 and 2.6.15 to address cache consistency issues.

Cheers,
Trond

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs