[Dovecot] Sieve error on Quota full (Dovecot 2.0.3)
Just switched my storage boxes to 2.0.3 with Pigeonhole. Noticing for those that have quota fulls, my syslog is getting full of errors from sieve:
dovecot: lda(user@host.com): Error: sieve: script /data/mail/host.com/user@host.com//.dovecot.sieve failed with unsuccessful implicit keep (user logfile /data/mail/host.com/user@host.com//.dovecot.sieve.log may reveal additional details)
Checking the .dovecot.sieve.log I see: sieve: info: started log at Sep 22 00:38:11. error: msgid=<dovecot-1285137490-378260-0@storagebox.host.com>: failed to store into mailbox 'INBOX': Quota exceeded ( mailbox for user is full).
Any way to suppress these quota full messages from filling up my logs and .dovecot.sieve.log ? I do like the .dovecot.sieve.log for debugging, but regular quota warnings seems excessive.
Thanks,
Cassidy
Op 22-9-2010 8:45, Cassidy Larson schreef:
Just switched my storage boxes to 2.0.3 with Pigeonhole. Noticing for those that have quota fulls, my syslog is getting full of errors from sieve:
dovecot: lda(user@host.com): Error: sieve: script /data/mail/host.com/user@host.com//.dovecot.sieve failed with unsuccessful implicit keep (user logfile /data/mail/host.com/user@host.com//.dovecot.sieve.log may reveal additional details)
Checking the .dovecot.sieve.log I see: sieve: info: started log at Sep 22 00:38:11. error: msgid=<dovecot-1285137490-378260-0@storagebox.host.com>: failed to store into mailbox 'INBOX': Quota exceeded ( mailbox for user is full).
Any way to suppress these quota full messages from filling up my logs and .dovecot.sieve.log ? I do like the .dovecot.sieve.log for debugging, but regular quota warnings seems excessive.
I'm wondering how your previous setup handled this. Currently, LDA and Pigeonhole don't maintain state information about persistent errors, causing a new error message in the logs for each failed delivery.
Regards,
Stephan.
sure enough... they did appear back in 1.2.x land. Sorry for the noise.
-c
I'm wondering how your previous setup handled this. Currently, LDA and Pigeonhole don't maintain state information about persistent errors, causing a new error message in the logs for each failed delivery.
Regards,
Stephan.
On Wed, 2010-09-22 at 12:17 +0200, Stephan Bosch wrote:
dovecot: lda(user@host.com): Error: sieve: script /data/mail/host.com/user@host.com//.dovecot.sieve failed with unsuccessful implicit keep (user logfile /data/mail/host.com/user@host.com//.dovecot.sieve.log may reveal additional details)
Checking the .dovecot.sieve.log I see: sieve: info: started log at Sep 22 00:38:11. error: msgid=<dovecot-1285137490-378260-0@storagebox.host.com>: failed to store into mailbox 'INBOX': Quota exceeded ( mailbox for user is full).
I don't think quota full should be logged as an error.. Without Sieve this is logged:
Info: msgid=unspecified: subject=: save failed to INBOX: Quota exceeded (mailbox for user is full) Info: msgid=unspecified: subject=: rejected: Quota exceeded (mailbox for user is full)
Also when this happens, does quota_full_tempfail setting work as expected?
I don't think quota full should be logged as an error.. Without Sieve this is logged:
Info: msgid=unspecified: subject=: save failed to INBOX: Quota exceeded (mailbox for user is full) Info: msgid=unspecified: subject=: rejected: Quota exceeded (mailbox for user is full)
Also when this happens, does quota_full_tempfail setting work as expected?
I haven't tried the quota_full_tempfail yet, because a lot of these are just accounts that haven't been used forever and have gotten full of spam and I dont know if I'd prefer tempfailing them to just bouncing.
Most of these accounts in question just have a simple .dovecot.sieve script generated by roundcube that only has "/* empty script */" in it. I assume that's why sieve is responding at all with a "failed with unsuccessful implicit keep".
Any chance sieve can just not worry about processing anything if there's nothing but comments in the .dovecot.sieve file? :)
-c
On 2010-09-22 8:53 PM, Cassidy Larson <alandaluz@gmail.com> wrote:
I haven't tried the quota_full_tempfail yet, because a lot of these are just accounts that haven't been used forever and have gotten full of spam and I dont know if I'd prefer tempfailing them to just bouncing.
Why waste the resources at all then? Just reject them at recipient validation stage with a custom "John no longer works here" message?
--
Best regards,
Charles
participants (4)
-
Cassidy Larson
-
Charles Marcus
-
Stephan Bosch
-
Timo Sirainen