Ah then, I believe the fix for this issue is part of  144489-10(kernel patch), related to CR BUGID 6976554, I could be wrong, we will reboot 1 of our servers with that on Thursday. (All others run IDR fix).

Lund


Ryan John wrote:
Hi Jorgen,

Thanks for the info.
It looks like this is my problem. I've seen the OpenOwner climb over 1 million, 
then I get NFS4ERR_STALE_CLIENTID errors.

In your earlier post, you said:
We can unmount/remount to free many locks, and about every 20 days, we
do maintenances where we restart nfsd.

I tried re-sharing just by doing a 'zfs set sharenfs=on pool/filesystem', but 
the locks kept increasing.
I restarted the nfs/server and the locks went down to zero.

The Nagios idea is good as well. I like graphs.

Regards
John

-----Original Message-----
From: [email protected] 
[mailto:[email protected]] On Behalf Of Jorgen Lundman
Sent: 22 March 2011 04:38
To: John Ryan
Cc: [email protected]
Subject: Re: [nfs-discuss] NFS4ERR_STALE_CLIENTID errors



I don't know anything specifically about your issue, nor do I work for Oracle,
but you could check to see if it is related to the trouble I was having:

http://opensolaris.org/jive/thread.jspa?threadID=118755

In particular, run the mdb command to see if you are also running out of
OpenOwner locks.

Lund


John Ryan wrote:
I see a lot of views, but no ideas?

I've reproduced this on another box, completely different hardware, but same OS.
I run around 16 simultaneous reads from 12 different clients, and after some 
time I get the same symptom.
Restarting nfs/server helped the first time I had this problem, but now it 
doesn't.
I've tried to do an svcadm restart nfs/server but for the last 30 minutes I still see the 
process "/sbin/sh /lib/svc/method/nfs-server stop 93"

Looks like I have to reboot.

Regards
John


--
Jorgen Lundman       | <[email protected]>
Unix Administrator   | +81 (0)3 -5456-2687 ext 1017 (work)
Shibuya-ku, Tokyo    | +81 (0)90-5578-8500          (cell)
Japan                | +81 (0)3 -3375-1767          (home)
_______________________________________________
nfs-discuss mailing list
[email protected]

Reply via email to