reveals little: rpc.lockd dies
Andy Sherman of At&T expressed a possible conflict
between 4.1 and 4.0.3 NFS servers which could have been the cause
of our problem.
Another cause could have been timeouts due to LAN
maintenance, i.e. ( an extended break in the cable )
----------------------
To: contel0!wizard!maxz
Subject: Re: rpc.lockd dies, processes hang with disk wait WHY?
In-Reply-To: Your message of Mon, 04 Feb 91 15:40:11 -0500.
<9102042040.AA00473@wizard.noname>
Date: Mon, 04 Feb 91 22:48:08 EST
Please state the releases of both client and server. Various
releases have all kinds of bugs (and patches) relating to
lock daemon crashes.
Andy Sherman/AT&T Bell Laboratories/Murray Hill, NJ
-------------
To: contel0!maxz@philabs.philips.com
Subject: Re: again: rpc.lockd dies, processes hang with disk wait WHY?
In-Reply-To: Your message of Tue, 05 Feb 91 10:43:10 -0500.
<9102051543.AA00881@wizard.noname>
Date: Tue, 05 Feb 91 17:42:16 EST
There is a patch for 4.0.3 to correct the "rpc.lockd core dumps"
problem. If you have software support, just call or write the
hotline. (1-800-USA-4SUN orhotline@sun.com).
*>Also, we have seen
*>strange interactions with 4.1 clients causing the lock daemon on a
*>4.0.3 server to die. I doubt there is a patch for that one yet, but
*>you might try giving the hotline a call.
Lock daemon problems are almost never your fault. The code has been
buggy in every release I've seen.
--
Andy Sherman/AT&T Bell Laboratories/Murray Hill, NJ
AUDIBLE: (201) 582-5928
READABLE:andys@ulysses.att.com or att!ulysses!andys
--------------------------------
From: Solbourne.COM!shirley@uunet (Shirley McAdams)
Message-Id: <9102051642.AA01277@Solbourne.COM>
To: contel0!maxz@philabs.philips.com
Subject: rpc.lockd dies, processes hang with disk wait
Mark,
Hi!! I work in Solbourne's National Technical Support Center and was
wondering what OS you are running and seeing this problem with?
4.0.3 or 4.1?
thanks,
Shirley McAdams UUCP: ..!{uunet,nbires,boulder}!stan!shirley
Longmont, CO 80501 Voice: 800-447-2861
--------------
Date: Tue, 5 Feb 91 12:00:05 MST
From: "Ric Anderson" <cs.arizona.edu!ric@uunet>
Message-Id: <9102051900.AA18791@cheltenham.cs.arizona.edu>
Received: by cheltenham.cs.arizona.edu; Tue, 5 Feb 91 12:00:05 MST
To: contel0!maxz@philabs.philips.com
Subject: rpc.lockd dies
There is a 4.1 vs. 4.0.3 problem here. I think the relevant patches
from Sun arePatch-ID# 100075-04
Keywords: lockd, rpc.lockd, rpc.statd file locking
Synopsis: lockd problems in 4.1 and 4.0.3
Date: 04/Oct/90
This patch is in /usr/SunPatches/Kernel/lockd.
Files changed: /vmunix, /usr/etc/rpc.lockd,
sys/{sun4,sun4c,sun3}/OBJ/klm_lockmgr.o, rpc.lockd, and ufs_lockf.o.
Hope this helps,
ric
------------------------------
Subject: again: rpc.lockd dies, processes hang with disk wait WHY?
Date: Tue, 05 Feb 91 10:43:10 EST
We are running on a mixed network of 4.1 and 4.0.3 systems, but the problem
with rpc.lockd dying happens between an NFS
server and client both running 4.0.3.
The server is a 4/280 16Mb RAM w/ 2 X 892Mb HDrives, the clients
have been both diskless 3/60s and diskful 4/60s.
The client systems are being used for software development and are configured
with kernels to include System V IPCisms.
Also the problem seems like it may be related to a portmapper failures,
seen if we attempt to halt the diskless 3/60s where the NFS server
is also the clients boot server-
Feb 4 10:12:30 berlioz login: ROOT LOGIN console
Feb 4 10:12:33 berlioz halt: halted by shutoff
Feb 4 10:12:35 berlioz syslogd: going down on signal 15
Feb 4 10:14:58 berlioz vmunix: Portmapper not responding; giving up
Feb 4 10:14:58 berlioz last message repeated 3 times
Feb 4 10:14:58 berlioz vmunix: syncing file systems... done
any insights ?
Thanx-
+ mark
