From: Michael Hale (mhale@transcomus.com)
Date: Thu Aug 30 2007 - 12:52:42 EDT
Yesterday, around 4:14pm one of our netra 440s (on a private network)
started to misbehave - I telneted to it and rather than a login
prompt, I got the error "STDIN is not a file" and then the session
would terminate.
I was able to console in and was seeing many messages on the
console like Jan 18 16:46:47 dare04 rpcbind: [ID 451547 daemon.error]
do_accept : can't open connection : Not enough space". Looking
further, I noticed the date was set to Jan 18th. I tried to run an
ls, which worked, but when I tried to run top, I got a /proc
filesystem full error message. I noticed that the date and time were
very wrong. I decided to reboot the host because it seemed confused
in a major way. After the host was up (filesystems all came up
clean), looking at dmesg, I saw the following error message: "Jan 1
00:03:56 dare04 genunix: [ID 125632 kern.warning] WARNING: Time-of-
day chip unresponsive; dead batteries?" and then the date jumped back
to being Jan 18 16:46:56 1913.
The following message "automountd[354]: [ID 618684 daemon.error]
cache_cleanup thread unexpected error 22" was being written to the
logs tens of thousands of times a second, for instance:
Jan 18 16:47:23 dare04 automountd[354]: [ID 618684 daemon.error]
cache_cleanup thread unexpected error 22
Jan 18 16:47:27 dare04 last message repeated 22865 times
Aug 29 22:13:55 davc02 hme: [ID 786680 kern.notice] SUNW,hme0 : No
response from Ethernet network : Link down -- cable problem?
Jan 18 16:47:27 dare04 automountd[354]: [ID 618684 daemon.error]
cache_cleanup thread unexpected error 22
Jan 18 16:47:38 dare04 last message repeated 65012 times
with sylog taking up 75% of the available CPU with a system load of
5. I killed automapd, which immediately brought down the load,
manually reset the date close enough to have xntpd take over, and
then restarted xntpd, at which time the time synced up and everything
seems to be working ok at this point in time.
So far, everything seems ok. I was wondering if anybody had seen
anything like this before? If the Time of Date battery does need to
be replaced, where is it on the 440 and how easy is it to replace?
Are there any other diagnostics I can run to check the health of the
system? I've never seen a solaris host freak out in this way before
and it has me worried.
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:42:16 EDT