[lug] Dual 1Ghz SMP Server Takes "Long Pause"
rriggs at doubleclick.net
Fri Nov 17 18:54:18 MST 2000
Nothing logged. No complaints on the serial console. It is all quite odd.
The lockups are happening for me when downloading and decoding
100+MB MIME email messages from a POP3 mailbox.
And I lied about the 866s working.. they seem to lock up for much less
time though. And I just found an interesting message in the syslog
immediately following a 20 minute stall that occured during testing
a few minutes ago:
"kernel: VM: killing process atd "
Looks like a VM problem. This may explain a few "lockups" we've
seen on a particular machine (the "favorite" of one of our power
users) in the past.
This sounds like there could be a nasty VM bug in the 2.2 kernel.
Sean Reifschneider wrote:
> On Fri, Nov 17, 2000 at 06:19:10PM -0700, Rob Riggs wrote:
> >I have a couple of Dell 2450 1Ghz SMP servers running Oct. KRUD.
> >Both seem to have the really odd problem of taking 40 minute pauses
> >when under high load and swapping. The servers respond to pings,
> Interestingly enough, we've been running into a situation with one
> of our clients where one of the mail machines in their cluster will
> hang up for about 20 minutes when sending a 150KB message to 100
> people. At least that seems to be the trigger... We're still
> investigating, but these are running something like KRUD 6.1 with
> 2.2.12 kernel on single processor boxes...
> Syslog on these boxes is going over the net, but we are getting some
> "out of memory" errors logged somwhere during this. You aren't seeing
> You think your Commodore 64 is really neato.
> What kinda chip you got in there, a Dorito? -- Weird Al
> Sean Reifschneider, Inimitably Superfluous <jafo at tummy.com>
> tummy.com - Linux Consulting since 1995. Qmail, KRUD, Firewalls, Python
> Web Page: http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
Unix System Administrator
DoubleClick/DARTmail - Broomfield, CO
More information about the LUG