Greetings -
I have migrated my Email server from an *old* OSX Server environment to
the following:
FreeBSD 14.3-RELEASE-p3
postfix-3.10.3,1
amavisd-new-2.13.1,1
Things looked good until this morning. I found the mail log file being
filled with error messages from amavis. Here is an excerpt:
Oct 4 04:52:26 franklin amavis[9876]: (09876-07) (!!)TROUBLE in
process_request: register_proc: BDB N c_get: BDB0061 PANIC: Invalid
argument, No such process. at
/usr/local/lib/perl5/site_perl/Amavis/DB/SNMP.pm line 224.
Oct 4 04:52:26 franklin amavis[9876]: (09876-07) (!)Requesting process
rundown after fatal error
Oct 4 04:52:26 franklin amavis[9876]: (09876-07) (!)TempDir removal:
tempdir is to be PRESERVED:
/var/amavis/tmp/amavis-20251004T030127-09876-RFXadT_r
Oct 4 04:52:26 franklin amavis[9876]: (09876-07) (!)Amavis::END: DB
unregistering failed:
Oct 4 04:52:26 franklin postfix/smtp[10895]: 1C942D7977:
to=<[email protected]>, relay=127.0.0.1[127.0.0.1]:10024,
delay=0.12, delays=0.02/0.03/0.01/0.06, dsn=4.3.2, status=deferred (host
127.0.0.1[127.0.0.1] said: 421 4.3.2 Service shutting down, closing
channel (in reply to end of DATA command))
Oct 4 04:52:26 franklin amavis[10897]: (!!)TROUBLE in child_init_hook:
BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY:
Fatal error, run database recovery, No such file or directory. at
/usr/local/lib/perl5/site_perl/Amavis/DB.pm line 82.
Oct 4 04:52:26 franklin amavis[10898]: (!!)TROUBLE in child_init_hook:
BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY:
Fatal error, run database recovery, No such file or directory. at
/usr/local/lib/perl5/site_perl/Amavis/DB.pm line 82.
The last line was repeating at about six errors per second, with an
incrementing PID. It had been doing this for over 3.5 hours I rebooted
the server, and this fixed the problem for a short while. It began
failing with the same symptom about 30 minutes later. The cause was
different, as shown by this excerpt:
Oct 4 09:24:54 franklin amavis[784]: (00784-11-3) Passed BAD-HEADER-4
{RelayedInbound,Quarantined}, [198.137.202.133]:37386 [198.137.202.133]
<[email protected]> ->
<[email protected]>, quarantine: badh-sIdXbs0a5POA, Queue-ID:
F1584D79BC, Message-ID:
<mailman.10522.1759587644.2093109.lede-comm...@lists.infradead.org>,
mail_id: sIdXbs0a5POA, Hits: -3.94, size: 7230, queued_as: 9DF49D7993,
dkim_sd=bombadil.20210309:lists.infradead.org, 771 ms
Oct 4 09:24:56 franklin amavis[784]: (00784-11-3) (!)Amavis::END: DB
unregistering failed:
Oct 4 09:24:56 franklin amavis[1420]: (!!)TROUBLE in child_init_hook:
BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY:
Fatal error, run database recovery, No such file or directory. at
/usr/local/lib/perl5/site_perl/Amavis/DB.pm line 82.
Oct 4 09:24:57 franklin amavis[1421]: (!!)TROUBLE in child_init_hook:
BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY:
Fatal error, run database recovery, No such file or directory. at
/usr/local/lib/perl5/site_perl/Amavis/DB.pm line 82.
This time I tried to simply restart amavis. I was able to stop it, but
the restart failed saying that the socket (port 10024) was in use.
Rebooting the server worked, and the system has been stable for the last
4 hours.
I have several questions related to this failure ...
1) I am surprised that amavisd-new doesn't recover from this failure.
It seems like the master process might detect the recurring failure and
restart. Is there something I can change in amavis that will improve this?
2) Given the error messages in the two excerpts, are there any hints on
what might have caused this?
3) Is there a better way to restart things? For example, restarting
postfix and then amavis?
Thanks in advance,
Paul
