[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Login failed, server replied: A000003 NO Maximum number of connections from user+IP exceeded ( mail_max_userip_connections )



On Tuesday 12 November 2013 22:40:00 Martin Steigerwald wrote:
> Am Dienstag, 12. November 2013, 22:21:17 schrieb Rainer Dorsch:
> > On Sunday 10 November 2013 14:10:56 Martin Steigerwald wrote:
> > > Am Sonntag, 10. November 2013, 13:58:49 schrieben Sie:
> > > > Am Sonntag 10.11.2013, 11:20:33 schrieb Martin Steigerwald:
> > > > > Note that it may not be actually a crashing resource, maybe just
> > > > > lots
> > > > > of
> > > > > new connections, would be nice if you can find out whether you
> > > > > akonadi
> > > > > imap resource gets restarted all the time (changing PID quickly) or
> > > > > its
> > > > > just new connection attempt as I could prove.
> > > > 
> > > > Wouldn't there be a popup notification for crashed akonadi resources?
> > > > I specifically remember one time when there were no popups but
> > > > something
> > > > akonadi-related kept downloading a couple hundred KB/s from my
> > > > university's
> > > > mail server. Re-starting akonadi didn't fix this, but stopping it
> > > > ended
> > > > the
> > > > download.
> > > > (And no, that wasn't a huge email/amount of emails being downloaded.
> > > > In
> > > > fact, there were no new emails.)
> > > > 
> > > > OTOH, I've also frequently had the "resource crashed" notification
> > > > spam
> > > > fill up my screen in 4.10, but haven't seen it in a while.
> > > 
> > > I am not sure. As described in bug report I saw a hint that akinadi pop
> > > resources crashed. But I don´t know how often.
> > > 
> > > Maybe its just reconnection attempts.
> > > 
> > > watch -n1 "ps aux | grep akonadi"
> > > 
> > > may give information about how PIDs change...
> > > 
> > > I didn´t do this back then and just watched port numbers changes with
> > > the
> > > netstat example I provided in bug report.
> > 
> > Hi Martin,
> > 
> > I just hit the problem again. PIDs are stable and I get lots of login
> > failed messages. Restarting dovecot cures the problem.
> > 
> > On the server side, I get lots of entries like this:
> > 
> > Nov 12 22:16:49 bokomoko dovecot: imap-login: Login: user=<user>,
> > method=PLAIN, rip=87.180.155.148, lip=aaa.bbb.ccc.ddd, TLS
> > Nov 12 22:16:49 bokomoko dovecot: imap-login: Disconnected (no auth
> > attempts): rip=87.180.155.148, lip=aaa.bbb.ccc.ddd, TLS: Disconnected
> 
> Thats good information to add to bug report.
> 
> Somewhere I read a hint it might be related to some TLS/SSL negotiation
> issue. Maybe it was in one of the the two bug reports.

I checked my server settings last week and changed in 
/etc/dovecot/dovecot.conf

575c575
<   #mail_max_userip_connections = 10
---
>   mail_max_userip_connections = 30

The problem did not reappear since then.

Is akonadi opening more parallel connections now? Is that expected?

Rainer

PS: If others have similar issues on virtual servers: watch out for the numfile 
limit on the server, this limits the number you can specify.

-- 
Rainer Dorsch
Lärchenstr. 6
D-72135 Dettenhausen
07157-734133
email: rdorsch@web.de
jabber: rdorsch@jabber.org
GPG Fingerprint: 5E6C 4AF9 363C 97A6 F820  C2E6 8CE0 5F33 07F3 EDC0
Full GPG key: http://pgp.mit.edu/

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: