From: Lyn St George (lyn_at_zolotek.net)
Date: Tue 07 Oct 2003 - 00:17:21 BST
Hallo
Over the last couple of weeks a number of vservers have been
triggering my monitoring system, because either services inside
a vserver become invisible to a 'ps ax', or the entire vserver seems
to have stopped when in fact it hasn't.
Example: 2 minutes ago I got an alert for a 4th vserver having gone
down, because a 'ps ax' from the host produced the line "x not running".
Entering the server and then doing a 'ps ax' produced this:
Server xxx is not running
ipv4root is now xxx
Host name is now xxx
New security context is 1008
[root_at_xxx /]ps ax
PID TTY STAT TIME COMMAND
1 ? S 5:37 init [3]
12225 pts/21 S 0:00 /bin/bash -login
12284 pts/21 R 0:00 ps ax
Despite this, everything is running, and ports 21, 22, 25, 80, 110 443 can
all be connected to.
What is plain is that the security context number is ridiculously high. On
a 'good' vserver this number is something like 26 or 32. On the 'bad' ones,
ie where 'ps ax' no longer shows the stuff that is running, this number is over
1000
What is going on?? Does anybody have any ideas at all, or has anyone
seen anything even remotely similar?
(kernel 2.4.18, vserver v12. Uptime 145 days with no such problems ever
over the last year or more. This host has 1Gb of memory, Athlon, current
load average: 0.75, 0.81, 0.72)
Any help will be greatly appreciated.
-
Cheers
Lyn St George
+---------------------------------------------------------------------------------
+ http://www.zolotek.net .. eCommerce hosting, consulting
+----------------------------------------------------------------------------------