looks like that are list-messages to myself, but not only because there are
much more messages until it crashes for different users

at least in case of the 3 last recent ones i got a message too

just a wild guess: "header_cache_readonly = yes" is AFAIK the only
recent change part of the game in case of lmtpd
__________________________________________________________________________________________

[root@mail:~]$ cat maillog | grep "message may be sent more than once" | wc -l
219

Apr  5 13:16:05 mail postfix/lmtp[17680]: 3g1FnK5Bkbz2W: 
to=<[email protected]>, orig_to=<[email protected]>,
relay=127.0.0.1[127.0.0.1]:24, delay=0.14, delays=0.03/0/0/0.11, dsn=4.4.2, 
status=deferred (lost connection with
127.0.0.1[127.0.0.1] while sending end of data -- message may be sent more than 
once)

Apr  4 13:14:31 mail postfix/lmtp[24043]: 3g0dnz2QkCz2g: 
to=<[email protected]>,
relay=127.0.0.1[127.0.0.1]:24, delay=0.19, delays=0.07/0/0/0.12, dsn=4.4.2, 
status=deferred (lost connection with
127.0.0.1[127.0.0.1] while sending end of data -- message may be sent more than 
once)

Apr  3 13:10:09 mail postfix/qmgr[736]: 3g01lP3TBzz23: 
from=<[email protected]>, size=4299,
nrcpt=1 (queue active)
Apr  3 13:10:09 mail postfix/lmtp[2823]: 3g01lP3TBzz23: 
to=<[email protected]>, relay=127.0.0.1[127.0.0.1]:24,
delay=0.15, delays=0.01/0/0/0.14, dsn=4.4.2, status=deferred (lost connection 
with 127.0.0.1[127.0.0.1] while
sending end of data -- message may be sent more than once)


Am 05.04.2014 16:44, schrieb Reindl Harald:
> Am 05.04.2014 15:47, schrieb Peter Milesson:
>> On 2014-04-05 13:54, Reindl Harald wrote:
>>> what i don't understand is why dbmail-lmtpd crashs once each day
>>> around the same time, even if i restart it at 3:00 AM, in case of
>>> imap/pop3 that could be a scheduled client always doing the same
>>> bad thing but hardly for LMTP
>>>
>>> dbmail-utils runs at a completly different time
>>>
>>> frankly i guess i could start strace with a cronjob between 13:00 and 14:00
>>> to a file with date/time because it's that predictable
>>>
>>> Apr  5 13:16:06 mail systemd[1]: dbmail-lmtpd.service holdoff time over, 
>>> scheduling restart.
>>> Apr  5 13:16:05 mail systemd[1]: Unit dbmail-lmtpd.service entered failed 
>>> state.
>>> Apr  4 13:14:32 mail systemd[1]: dbmail-lmtpd.service holdoff time over, 
>>> scheduling restart.
>>> Apr  4 13:14:31 mail systemd[1]: Unit dbmail-lmtpd.service entered failed 
>>> state.
>>>
>> Hi Harald,
>>
>> Could it be that somebody is sending some kind of news letter, report, etc 
>> at the same time every day? Buffer
>> overrun? Something happening due to logrotate or cron?
> 
> i will dig through the maillog for messages received at the 
> crash-timestamps.......
> 
> maybe, on the otherside it should fail more than once since postfix would 
> queue
> and retry - but *wait*: what if that messages are not stored correctly and
> dbmail-lmtpd answers to postfix with a OK status code before it segfaults?
> 
> if that guess is true it would be a serious bug because the "OK" status should
> be only returned to postfix if the complete message is stored without errors
> 
> i fear that explains the dbmail-utils below errors too:
>  * 8 crashes
>  * [8] missing rfcsize values
>  * [8] missing envelope values
>  * [8] un-cached physmessages
> 
> 2014-04-05 13:16:06   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-04-04 13:14:32   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-04-03 13:10:10   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-04-02 13:07:33   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-04-01 13:01:13   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-03-31 13:00:10   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-03-30 13:13:38   dbmail-lmtpd.service holdoff time over, scheduling 
> restart.
> 2014-03-30 12:56:31   dbmail-lmtpd.service holdoff time over, scheduling 
> restart
> _________________________________________________________________
> 
> 02.04.2014
> 
> [root@mail:~]$ /usr/sbin/dbmail-util -y -c -t -u -b -p -d -r 24h
> Ok. Found [4] missing rfcsize values.
> EEEEOk. Found [4] missing envelope values.
> EEEEOk. Found [4] un-cached physmessages.
> _________________________________________________________________
> 
> 05.04.2014
> 
> Repairing DBMAIL for rfcsize field...
> Ok. Found [8] missing rfcsize values.
> EEEEEEEE
> Repairing DBMAIL for cached envelopes...
> Ok. Found [8] missing envelope values.
> EEEEEEEE
> Repairing DBMAIL for cached header values...
> Ok. Found [8] un-cached physmessages.
> EEEEEEEE

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
DBmail mailing list
[email protected]
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Reply via email to