[Kolab-devel] Massive loss of events in enterprise35 client

ITSEF Admin itsef-admin at brightsight.com
Wed Jan 14 14:38:48 CET 2009


Hi all,

I'm Cc'ing this to both lists, as  I regard this as a major problem - "devel" 
to get some input from the developers and "users" just to see whether anybody 
else has seen things like this anywhere.

For the first time in enterprise35 (that I know of in our company), the old, 
ugly problem rears its head again: Massive loss of events from one's agenda. 
This time it has happened to myself - yesterday evening, several hundred 
events were deleted from my agenda, without my intervention. Here's the 
forensics:
- kontact --version
  Qt: 3.3.8b
  KDE: 3.5.10
  Kontact: 1.2.9 (enterprise35 20081202.891698)
  (Kubuntu 8.04)
- $HOME is on NFS
- Kontact was running, but I do not recall actually working with Kontact at 
the time, though I did start a new mail shortly afterwards
- According to the imap server logs (Kolab 2.1), the events in question were 
deleted - I can see hunderds of lines like this:
imapd.log.0:Jan 13 18:13:19 HOST <debug> imaps[3259]: mailbox_expunge: 
removing mail brightsight.com!user.USER.Calendar:18646
- Earlier that day, I had a crash - most likely due to connection problems 
with our mailserver (we had firewall problems), but I had already restarted 
Kontact
- I did not find anything suspicious in .xsession-errors, except for a bunch 
of SSL errors I could not place:
5594:error:140940E5:lib(20):func(148):reason(229):s3_pkt.c:990:
5594:error:14094415:lib(20):func(148):reason(1045):s3_pkt.c:1053:SSL alert 
number 45
- I am using four different calendars - my own, another "single user" 
calendar, one group calendar and a shared folder calendar. Only my own 
calendar was affected.
- Oddly enough, not all of the events in my calendar were deleted - 9 messages 
were left behind on the server, one of which was a real event from last 
October. The other eight were empty.


Another user also lost all events yesterday, however, in his case it happened 
with the Kontact instance that was still running during the connection 
problems with the mail server. His events were deleted (again, according to 
imapd.log) about three hours after the connection had been restored. He had 
not restarted Kontact since. His .xession-errors has already been 
overwritten, so no help there.


We have experienced problems like this rather frequently with proko2. 
enterprise35 is doing much better (in fact, I was lulled into the hope that 
the problem might have vanished) - but apparently, there's still something 
not quite right...

Any input - as always - welcome!

Regards,

Thomas




More information about the devel mailing list