1 |
I have (the latest) spamassassin 3.1.0 (on Perl 5.8.7) installed on my |
2 |
not-extremely-powerful small-scale Gentoo server (P2-450 with 256Mb RAM |
3 |
and 500Mb swap - not running X) and I'm experiencing a recurring |
4 |
fault... I'm running spamd and using the spamassassin client to |
5 |
re-direct email to it from my systemwide procmail setup (using the |
6 |
recommended | /usr/bin/spamc -f approach with a size limiter to avoid |
7 |
over-burdening spamassassin with any massive emails.) |
8 |
|
9 |
Everything seems to work well for a fair while, then the spamd process |
10 |
gets "wedged" - and in spite of there being a near-0 load, and dozens |
11 |
and dozens of messages reported by mailq, none of my mail gets processed |
12 |
in a hurry - and after a long while (maybe every half-an-hour, say) an |
13 |
email is delivered from the head of my mail queue - which has not passed |
14 |
through spamd (according to the headers) - and has no spam-score attached. |
15 |
|
16 |
I read some suggestions a long while ago which said that "--round-robin" |
17 |
as an option was a work-around for a bug with the same consequences. |
18 |
My /etc/conf.d/spamd currently has the options: |
19 |
|
20 |
SPAMD_OPTS="-m 5 -c -H -l --round-robin" |
21 |
|
22 |
However, I still get this problem. I thought that it was related to the |
23 |
length of time the spamd process had been running - so I set a cron job |
24 |
to re-start the server at an unusual time early every morning... but |
25 |
this hasn't been an successful work around either. It now seems that to |
26 |
be triggered by an increased system load - today I ran |
27 |
|
28 |
# emerge apache squirrelmail |
29 |
|
30 |
then left... when I returned the emerge had long-since completed by my |
31 |
spamd remained "stuck." |
32 |
|
33 |
Is this a problem everyone is having? I had no problem like this with |
34 |
the elder (3.0.4) version of spamassassin I had installed previously on |
35 |
the same hardware. |
36 |
-- |
37 |
gentoo-user@g.o mailing list |