Coda File System

codasrv gets stuck

From: Steve Simitzis <steve_at_saturn5.com>
Date: Tue, 2 Dec 2003 08:25:05 -0800
after several months of flawless operation, codasrv has decided to
behave oddly again. this has happened twice in the last two days, so
i'm not prepared to dismiss it as a fluke.

the problem is that codasrv will freeze, apparently unbind all its
connections, and refuse to do much of anything. the only way to get it
running again is to kill -9 codasrv, and restart everything.

what's curious is that before the freeze, i see several of these:

07:23:44 ****** WARNING entry at 0x8188a18 already has deqing set!

until all the connections are dropped:

07:25:00 Worker2: Unbinding RPC connection 3289
07:25:00 Worker2: Unbinding RPC connection 4364
07:25:00 Worker2: Unbinding RPC connection 11512

i ran gdb against codasrv, and here's where it said it was:

(gdb) where
#0  0x420e187e in select () from /lib/i686/libc.so.6
#1  0x4008a28c in __DTOR_END__ () from /usr/lib/liblwp.so.2
#2  0x400860d9 in IOMGR (dummy=0x0) at iomgr.c:356
#3  0x40087f56 in Create_Process_Part2 () at lwp.c:796
(gdb) quit


the server is running coda-server-6.0.3 on linux 2.4.20 (redhat 7.3)
rpc2-1.20, lwp-1.10.

my coda clients are all running coda-client-6.0.2 on linux 2.4.22 (redhat 8)
rpc2-1.19, lwp-1.10.

-- 

steve simitzis : /sim' - i - jees/
          pala : saturn5 productions
 www.steve.org : 415.282.9979
  hath the daemon spawn no fire?
Received on 2003-12-02 11:28:52