.iwmnt cache not updating

I have some interesting cache behaviour occuring. If I preview a particular image that I have just imported (to over-write an existing one), the older existing one is the one that is displayed NOT the one that I have just imported.

A look on the filesystem reveals:

#ls -las /.iwmnt/default/main/NRM/WORKAREA/webdev/waterwise
-rw-rw-r--   1 gliddonl intranet   14480 Feb 14 15:18 waterwise_topic_banner.jpg

#ls -las /iwmnt/default/main/NRM/WORKAREA/webdev/waterwise
-rw-rw-r--   1 robinsct intranet   16267 Feb 18 11:48 waterwise_topic_banner.jpg

Notice that the cached version (.iwmnt) is the older image, whilst the non-cached version (iwmnt) is the more recent imported image.

I have undertaken a iwfreeze +0 command as per KB article 1344. This did not make any difference.

I have been having some recurring "Stale NFS handle" errors recently (not sure why)....I wonder if it related? Any help is appreciated Smiley Very Happy

Find more posts tagged with

Comments

Adam Stoller

I have some interesting cache behaviour occuring. If I preview a particular image that I have just imported (to over-write an existing one), the older existing one is the one that is displayed NOT the one that I have just imported.

A look on the filesystem reveals:
#ls -las /.iwmnt/default/main/NRM/WORKAREA/webdev/waterwise
-rw-rw-r--   1 gliddonl intranet   14480 Feb 14 15:18 waterwise_topic_banner.jpg

#ls -las /iwmnt/default/main/NRM/WORKAREA/webdev/waterwise
-rw-rw-r--   1 robinsct intranet   16267 Feb 18 11:48 waterwise_topic_banner.jpg
Notice that the cached version (.iwmnt) is the older image, whilst the non-cached version (iwmnt) is the more recent imported image.

I have undertaken a iwfreeze +0 command as per KB article 1344. This did not make any difference.

I have been having some recurring "Stale NFS handle" errors recently (not sure why)....I wonder if it related? Any help is appreciated .

You have your cached and non-cached mount points mixed up - .iwmnt is non-cached -- the fact that the non-cached mount point appears to be messed up would suggest that the stale NFS handles probably have something to do with it. Have you tried physically rebooting your server?

brizrobbo

Oops...thanks for that. I think you're right about the stale NFS handles. I've been periodically getting "Stale NFS filehandle" errors for a little while now. Typically, a reboot of Teamsite (not the whole server) fixes it...for a time.

I actually had another thread posted about this, and was directed to the KB which indicated that this issue could be caused by TS not shutting down properly (and the NFS shares not unmounting due to some activity). The thing is, is that whenever we now shutdown down TS we definitely make sure that the NFS shares are unmounted prior to starting up. So I'm curious as to what might be causing this!!

Would you actually suggest a physical reboot of the server?

Adam Stoller

Oops...thanks for that. I think you're right about the stale NFS handles. I've been periodically getting "Stale NFS filehandle" errors for a little while now. Typically, a reboot of Teamsite (not the whole server) fixes it...for a time.

I actually had another thread posted about this, and was directed to the KB which indicated that this issue could be caused by TS not shutting down properly (and the NFS shares not unmounting due to some activity). The thing is, is that whenever we now shutdown down TS we definitely make sure that the NFS shares are unmounted prior to starting up. So I'm curious as to what might be causing this!!

Would you actually suggest a physical reboot of the server?

Although far less common a need on Unix than on Windows, an occasional reboot of the actual host server is not a bad thing - and if you're getting stale NFS handles, probably the best thing to do. If the problem persists with the stale NFS handles after the reboot, then you have to do more investigation as to what other process(es) might be causing the mount points to become disconnected and/or communication with them disrupted.

cesarecastro

Ghoti,

A reboot should only be your last resource. Especially if Teamsite is running on a server with more then one application running on it. If I were robo, I would contact support and have them help with the issue. If there is a bug with Teamsite this will also ensure that it is documented and fix asap or on the next patch. Or maybe robo has not kept TS uptodate with the latest patches. But a reebot as a fix to this type of issue for an enterprise application would not be acceptable. In any case, these are just my thought on this.

Cesar

Adam Stoller

Ghoti,

A reboot should only be your last resource. Especially if Teamsite is running on a server with more then one application running on it. If I were robo, I would contact support and have them help with the issue. If there is a bug with Teamsite this will also ensure that it is documented and fix asap or on the next patch. Or maybe robo has not kept TS uptodate with the latest patches. But a reebot as a fix to this type of issue for an enterprise application would not be acceptable. In any case, these are just my thought on this.

Cesar

If you're dealing with Stale NFS handles - it's usually because one ore more mounted partitions went off-line (probably not gracefully) and the server is still holding on to them in one way or another.

If stopping and starting the TeamSite service doesn't fix the problem - a reboot of the server as a whole is not beyond reason.

If the problem persists - I'd say the local system and network administrators are probably the ones to help resolve the problem, as it is unlikely to be a bug in TeamSite (though TeamSite might be the primary application that is affected by it).

brizrobbo

I've scheduled a restart of the server for tomorrow AM. I'll let you all know how it goes.

brizrobbo

Just an update....

Rebooted the server this morning. And manually brought up all the relevant processes (iwserver, iwod60, etc.). I say manually, insofar as we normally script the start of the Interwoven products, and I wanted to remove that as a source of our NFS problems. So, we started everything in order and confirmed successful mountings etc.

Everything looks sweet so far Smiley Happy

. NFS problems have been resolved....so, let's hope it stays that way. If not, then obviously some process is intefering with NFS, and we'll have to dig deeper (no doubt with the help of IWOV support).

If there is anything to add over the next week or so... I will do so.