E-Mail catchall is lost (all messages bounce), requires re-config

[aegide]

Hi,

this is

Linux my.server.net 3.2.0-32-generic #51-Ubuntu SMP Wed Sep 26 21:33:09 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

running

11.0.9 Ubuntu 12.04 110120608.16

with all available microupdates on air.


I've repeatedly lost catchall forwarders for multiple domains, e.g.

domain my.server.net doesn't have any users/email accounts configured but an overall

- email enabled for this domain
- forward all unknown emails to 'email@otherserver.net'
- ...

Postfix log states

Mar 4 11:39:55 my.server.net postfix/smtpd[17731]: connect from mail-la0-f44.google.com[209.85.215.44]
Mar 4 11:39:55 my.server.net postfix/smtpd[17731]: NOQUEUE: reject: RCPT from mail-la0-f44.google.com[209.85.215.44]: 550 5.1.1 <someadress@my.server.net>: Recipient address rejected: User unknown in virtual mailbox table; from=<someone@somewhere.com> to=<someadress@my.server.net> proto=ESMTP helo=<mail-la0-f44.google.com>
Mar 4 11:39:55 my.server.net postfix/smtpd[17731]: disconnect from mail-la0-f44.google.com[209.85.215.44]

before simply hitting "OK" again in the plesk panel configuration (no changes made), after which plesk restarts named and reconfigures virtual hash table:

Mar 4 11:39:55 my.server.net /usr/lib/plesk-9.0/psa-pc-remote[17149]: Message aborted.
Mar 4 11:41:29 my.server.net named[914]: received SIGHUP signal to reload zones
Mar 4 11:41:29 my.server.net named[914]: loading configuration from '/etc/named.conf'
Mar 4 11:41:29 my.server.net named[914]: using default UDP/IPv4 port range: [1024, 65535]
Mar 4 11:41:29 my.server.net named[914]: using default UDP/IPv6 port range: [1024, 65535]
Mar 4 11:41:29 my.server.net named[914]: sizing zone task pool based on 9 zones
Mar 4 11:41:29 my.server.net named[914]: Warning: 'empty-zones-enable/disable-empty-zone' not set: disabling RFC 1918 empty zones
Mar 4 11:41:29 my.server.net named[914]: reloading configuration succeeded
Mar 4 11:41:29 my.server.net named[914]: reloading zones succeeded
Mar 4 11:41:29 my.server.net named[914]: zone 246.214.85.in-addr.arpa/IN: loaded serial 1362393689
Mar 4 11:41:29 my.server.net named[914]: zone my.server.net/IN: loaded serial 1362393688
Mar 4 11:41:36 my.server.net postfix/trivial-rewrite[17735]: table hash:/var/spool/postfix/plesk/virtual(0,lock|fold_fix) has changed -- restarting

after which mail comes through again:

Mar 4 11:41:50 my.server.net postfix/smtpd[17830]: connect from mail-la0-f53.google.com[209.85.215.53]
Mar 4 11:41:50 my.server.net postfix/smtpd[17830]: 7C69FDFBA0: client=mail-la0-f53.google.com[209.85.215.53]
Mar 4 11:41:50 my.server.net postfix/cleanup[17834]: 7C69FDFBA0: message-id=<CAN3PQ3YpyxVg3stWip2hnyVLewt+ibZFVKqyTDXOcnHG7sQ@mail.gmail.com>
Mar 4 11:41:50 my.server.net /usr/lib/plesk-9.0/psa-pc-remote[17149]: handlers_stderr: SKIP
Mar 4 11:41:50 my.server.net /usr/lib/plesk-9.0/psa-pc-remote[17149]: SKIP during call 'check-quota' handler
Mar 4 11:41:50 my.server.net postfix/qmgr[17361]: 7C69FDFBA0: from=<someone@somewhere.com>, size=1643, nrcpt=1 (queue active)
Mar 4 11:41:50 my.server.net postfix/smtpd[17830]: disconnect from mail-la0-f53.google.com[209.85.215.53]
Mar 4 11:41:50 my.server.net postfix/smtp[17837]: 7C69FDFBA0: to=<catchallfwdfrom-myservernet@other.de>, orig_to=<blah@my.server.net>, relay=imap.other.de[IP]:25, delay=0.46, delays=0.25/0.03/0.15/0.04, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as E05DB1AF9)
Mar 4 11:41:50 my.server.net postfix/qmgr[17361]: 7C69FDFBA0: removed

Any ideas? Additional log needs?

If there's no quick fix, how can i reload/reset configuration from the commad line? I would check for a bounce from another server and in case of bounce reload config on the server running plesk that lost the catchall config.

Thanks!

aeg

 

[aegide]

Hi,

it happened again today. All i did was add a couple of domains, email fwds, catchalls and websites. No server reconfiguration/ update/ ... on the shell whatsoever.

Postmap output after loosing the catchall:

postmap -s hash:/var/spool/postfix/plesk/virtual | grep ^\@
@a.de someone@somewhere.de
@b.c.name someone@somewhere.de
@d.e.de someone@somewhere.de
@f.name someone@somewhere.de
@g.de someone@somewhere.de
@h.de someone@somewhere.de
@i.de someone@somewhere.de
@j.de someone@somewhere.de

While simply going back to Domainconfig->Email->Change Configuration and Clicking "OK" (the catchall is still there, combo checked) leads to

postmap -s hash:/var/spool/postfix/plesk/virtual | grep ^\@
@a.de someone@somewhere.de
@b.c.name someone@somewhere.de
@d.e.de someone@somewhere.de
@x.z.name someone@somewhere.de
@f.name someone@somewhere.de
@g.de someone@somewhere.de
@h.de someone@somewhere.de
@i.de someone@somewhere.de
@j.de someone@somewhere.de

being back in.

How can i debug this any further? It seems pretty random as i had previously checked the hashtable in between pleskpanel actions w/o any apparent changes made to it and then suddenly the entry (or another for that matter) is missing.

what's going on?

aeg

ps: will setup inode ctime watch on the file with diffing and auto notification upon change for debugging purposes but would really like some insight into this in the meantime. thanx.