Hard Lock

[Glenn Steen]

    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "US-ASCII" character set.  ]
    [ Some characters may be displayed incorrectly. ]

On 10/01/06, Information Services <lbcadmin at gmail.com> wrote:
> Once again one of my mailscanner systems locked up.  I had quite a scare
> with this one, and I am sure I will be in during the weekend building
> another system for when this one kills over.  This time I had to take out
> the RAM and replug the HDs in order to it to get a sense of itself, but I am
> wondering if it has something to do with the powerleaps we have in the Dell
> Optiplex G1s.  Could they be overheating and causing the lockups?  If this
> is possible that would explain why it booted after I messed with the other
> hardware.  But this is what I retried from the /var/log/messages.  Better
> information, but not something I understand.  Can you make sense of it??

Yes, definitely. If you have a budget, try getting new boxes.... In my
experience it is very seldom cost-effective to muck about with CPU
upgrades.... When you do them, the rest of the system is getting
old.... and even though the system _should_ work OK with the new
parts, *something* almost always has "gone marginal". Do as Scott says
and revert at least this box to the original CPUs, or just replace it
with something new and shiny.... Opteron systems are getting _cheap_
these days;-).

>  ---------------------------
>  Jan 10 13:05:03 wks-lin9 crond(pam_unix)[11999]: session closed for user
> root
> Jan 10 13:05:15 wks-lin9 crond(pam_unix)[12000]: session closed for user
> root
> Jan 10 13:05:22 wks-lin9 crond(pam_unix)[11473]: session closed for user
> root
>
> Jan 10 13:06:49 wks-lin9 smbd[12451]: [2006/01/10 13:06:49, 0]
> lib/util_sock.c:get_peer_addr(1000)
> Jan 10 13:06:49 wks-lin9 smbd[12451]: getpeername failed. Error was
> Transport endpoint is not connected
> Jan 10 13:06:49 wks-lin9 smbd[12451]: [2006/01/10 13:06:49, 0]
> lib/util_sock.c:get_peer_addr(1000)
>
> Jan 10 13:06:49 wks-lin9 smbd[12451]: getpeername failed. Error was
> Transport endpoint is not connected
> Jan 10 13:06:49 wks-lin9 smbd[12451]: [2006/01/10 13:06:49, 0]
> lib/util_sock.c:write_socket_data(430)
> Jan 10 13:06:49 wks-lin9 smbd[12451]: write_socket_data: write failure.
> Error = Connection reset by peer
>
> Jan 10 13:06:49 wks-lin9 smbd[12451]: [2006/01/10 13:06:49, 0]
> lib/util_sock.c:write_socket(455)
> Jan 10 13:06:49 wks-lin9 smbd[12451]: write_socket: Error writing 4 bytes to
> socket 24: ERRNO = Connection reset by peer
>
> Jan 10 13:06:49 wks-lin9 smbd[12451]: [2006/01/10 13:06:49, 0]
> lib/util_sock.c:send_smb(647)
> Jan 10 13:06:49 wks-lin9 smbd[12451]: Error writing 4 bytes to client. -1.
> (Connection reset by peer)
> Jan 10 13:06:50 wks-lin9 winbindd[2518]: [2006/01/10 13:06:50, 0]
> tdb/tdbutil.c:tdb_log(725)
>
> Jan 10 13:06:50 wks-lin9 winbindd[2518]:
> tdb(/var/cache/samba/netsamlogon_cache.tdb): rec_free_read
> bad magic 0x42424242 at offset=7252
> -------------------
>
> that is the last information posted to messages before the boot up
> information is written.
>

Interrestingly enough, this is "microsoft networking" (samba)
complaining loudly on the inability to resolve names to addresses,
because the "nameserver" (winbindd) died... and that seem to have died
from a database corruption. Could mean nothing, of course, just show
that these died first. Most likely show a problem with your CPU, RAM
or HDDs (likely contoller promlems in the third case), with CPU and
RAM being most likely culprits.
standard measures apply.... Swap _ONE_ thing at a time and see if that
helps... Although, with intermittent hangs like this, it's often very
hard to be sure if some action has solved the problem.
Simlest solution is to get brand  new HW.

--
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the Wiki (http://wiki.mailscanner.info/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).

Support MailScanner development - buy the book off the website!