Sudden instability

[Todd Lyons]

Up until yesterday, the mail system was very stable (3 load balanced
SMTP's handling about 50K messages each per day with clamav and
spamassassin).  Yesterday I made one config change to spamd, telling it
not to convert marked spams to an attachment (report_safe 0).  SA marks
it at 5.0 and the milter rejects it at 7.0.  These are RedHat 9.0 boxen.

I didn't notice it until last night, but the spamass-milter process is
now dying on me at fairly regular intervals.  Last night it was about
once per hour.  I have a cron job that checks status every 15 minutes
and restarts if necessary.

Looking through the archives, one comment by Dan *REALLY* stood out.
"Standard behaviour for threaded apps under Linux when they crash.  I
bet if you run 'ps axw | grep spamass-milter', you see one process
instead of the usual 3."  This is what I am seeing now too.  I'm not
sure what changed on the three machines from yesterday until today other
than some whitelists and the report_safe option.  I cannot say with
certainty that there was three processes yesterday before I changed the
report_safe option.  I tend to think that the report_safe option is a
red herring, but maybe not.

There have been no new packages installed since I installed clamav-0.70
back on June 1.

My uptime is around 100 days, so it's not been rebooted for some time.
I can't see how that would affect it either.  These machines each have 2
Gigs of RAM, so they're not starved of resources.  Load average usually
is below 4.0 for each of the machines though it does spike
occassionally.

address@hidden log]# ps ax | grep spamass-milter | grep -v grep
27722 pts/1    S      0:00 spamass-milter -p /var/run/spamass.sock -f -r 7
address@hidden log]# ldd `which spamass-milter`
        libstdc++.so.5 => /usr/lib/libstdc++.so.5 (0x4001c000)
        libm.so.6 => /lib/tls/libm.so.6 (0x400cf000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x400f1000)
        libpthread.so.0 => /lib/tls/libpthread.so.0 (0x400fa000)
        libc.so.6 => /lib/tls/libc.so.6 (0x42000000)
        /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)

Any suggestions or comments are appreciated.

I will probably have to switch to CVS code because I need the -e option.
We have virtual users and I need to have per user spamassassin
configuration control.
-- 
Regards...              Todd
They that can give up essential liberty to obtain a little temporary 
safety deserve neither liberty nor safety.       --Benjamin Franklin
Linux kernel 2.6.3-8mdkenterprise   2 users,  load average: 0.04, 0.04, 0.00