You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Will Yardley <sa...@veggiechinese.net> on 2004/09/24 19:15:01 UTC
spamd dying?
Has anyone else seen a problem w/ spamd dying sometimes (after working
for a while)? I have been seeing this in the 3.0 rcs. I'm about to
upgrade to 3.0 release, but I'm wondering if anyone else has seen this.
Much more detailed information in bug #3667 (bugzilla.spamassassin.org)
Re: spamd dying?
Posted by Will Yardley <sa...@veggiechinese.net>.
Oh - also, sorry for the self-followup...
normal ps output:
root 8978 1 0 17:55 ? 00:00:00 /usr/bin/perl -T -w /usr/local/bin/spamd -d -D
root 8979 8978 0 17:55 ? 00:00:00 spamd child
root 8980 8978 0 17:55 ? 00:00:00 spamd child
root 8981 8978 0 17:55 ? 00:00:00 spamd child
root 8982 8978 0 17:55 ? 00:00:00 spamd child
root 8983 8978 0 17:55 ? 00:00:00 spamd child
ps output after spamd seems to die - note how there seems to be one
process left that's a child of the master process.
root 21117 1 0 Oct11 ? 00:01:38 /usr/bin/perl -T -w /usr/local/bin/spamd -d -D
root 6102 21117 0 17:51 ? 00:00:00 /usr/bin/perl -T -w /usr/local/bin/spamd -d -D
Re: spamd dying?
Posted by Will Yardley <sa...@veggiechinese.net>.
On Tue, Sep 28, 2004 at 11:56:52AM -0700, Justin Mason wrote:
> Will Yardley writes:
> > On Fri, Sep 24, 2004 at 10:22:01AM -0700, Justin Mason wrote:
> > > Will Yardley writes:
> > > > Has anyone else seen a problem w/ spamd dying sometimes (after
> > > > working for a while)? I have been seeing this in the 3.0 rcs.
> > > > I'm about to upgrade to 3.0 release, but I'm wondering if anyone
> > > > else has seen this.
> > > > Much more detailed information in bug #3667
> > > I'm almost certain it's a fixed bug.
> > I don't have any more useful information yet (strace output), but I
> > /am/ seeing this with SA 3.0 release.
> > What will provide the most useful information for debugging this?
> > Should I run the daemon in debug mode, or just run strace on it and
> > see what happens when it dies?
> *both* would provide info that will help. ;)
The strace thing didn't work out, because it takes so long for spamd to
die, and the strace output is (of course) quite verbose, at least if I
follow child processes. I'm assuming that running strace on the master
processes (which don't seem to die) won't help much.
Spamd has still been dying pretty consistently on me - takes a few days
to happen. This is the last reference I see to spamd in the logs...
didn't notice anything unusual.
....
Oct 12 06:22:27 drama spamd[21120]: logmsg: identified spam (10.4/2.7) for dallas:10000 in 4.3 seconds, 3969 bytes.
Oct 12 06:22:27 drama spamd[21120]: identified spam (10.4/2.7) for dallas:10000 in 4.3 seconds, 3969 bytes.
Oct 12 06:22:27 drama spamd[21120]: logmsg: result: Y 10 - BAYES_99,DNS_FROM_AHBL_RHSBL,HTML_50_60,HTML_IMAGE_ONLY_16,HTML_IMAGE_RATIO_02,HTML_MESSAGE,HTML_TEXT_AFTER_BODY,HTML_TEXT_AFTER_HTML,RCVD_IN_BL_SPAMCOP_NET,RCVD_IN_SBL,SPF_HELO_FAIL,URIBL_OB_SURBL,URIBL_SBL,URIBL_WS_SURBL scantime=4.3,size=3969,mid=<92...@zmawidknoZ-zwhhwn>,bayes=1,autolearn=no
Oct 12 06:22:27 drama spamd[21120]: result: Y 10 - BAYES_99,DNS_FROM_AHBL_RHSBL,HTML_50_60,HTML_IMAGE_ONLY_16,HTML_IMAGE_RATIO_02,HTML_MESSAGE,HTML_TEXT_AFTER_BODY,HTML_TEXT_AFTER_HTML,RCVD_IN_BL_SPAMCOP_NET,RCVD_IN_SBL,SPF_HELO_FAIL,URIBL_OB_SURBL,URIBL_SBL,URIBL_WS_SURBL scantime=4.3,size=3969,mid=<92...@zmawidknoZ-zwhhwn>,bayes=1,autolearn=no
(message then appears to be successfully delivered to the user).
After this, there are no spamd references in the log until I restart
spamd.
Re: spamd dying?
Posted by Will Yardley <sa...@veggiechinese.net>.
On Tue, Sep 28, 2004 at 11:56:52AM -0700, Justin Mason wrote:
> Will Yardley writes:
> > On Fri, Sep 24, 2004 at 10:22:01AM -0700, Justin Mason wrote:
> > > Will Yardley writes:
> > > > Much more detailed information in bug #3667 (bugzilla.spamassassin.org)
> > > I'm almost certain it's a fixed bug.
> > What will provide the most useful information for debugging this? Should
> > I run the daemon in debug mode, or just run strace on it and see what
> > happens when it dies?
> *both* would provide info that will help. ;)
So just do this?
# /usr/local/bin/spamd -d -D 2> /tmp/spamd.debug.out
or do you have to do something special to get debug to work when running
as a daemon? Do I need to do something to get the child processes to
spit out debug information too?
for strace, I'm doing:
# strace -f -p[pid] 2> /tmp/spamd.strace.out &
# disown %1
(-f should be chasing the child processes as well)
so far here I'm just seeing:
pause(
in the strace output (I'd expect much more verbose output here), and
just:
creating INET socket:
Listen: 128
LocalAddr: 127.0.0.1
LocalPort: 783
Proto: 6
ReuseAddr: 1
Type: 1
debug: SpamAssassin version 3.0.0
debug: Score set 0 chosen.
debug: Storable module v1.014 found
in the debug file.
Re: spamd dying?
Posted by Will Yardley <sa...@veggiechinese.net>.
On Fri, Sep 24, 2004 at 10:22:01AM -0700, Justin Mason wrote:
> Will Yardley writes:
> > Has anyone else seen a problem w/ spamd dying sometimes (after working
> > for a while)? I have been seeing this in the 3.0 rcs. I'm about to
> > upgrade to 3.0 release, but I'm wondering if anyone else has seen this.
> >
> > Much more detailed information in bug #3667 (bugzilla.spamassassin.org)
>
> I'm almost certain it's a fixed bug.
I don't have any more useful information yet (strace output), but I /am/
seeing this with SA 3.0 release.
What will provide the most useful information for debugging this? Should
I run the daemon in debug mode, or just run strace on it and see what
happens when it dies?