Re: [Evolution] Calendar-EWS generates zombies [evolution] <defunct>

[Michael Felsberg <michael felsberg liu se>]

> > > Fontconfig error: "/etc/fonts/conf.d/65-khmer.conf", line 14: out of
> > > memory
> > > Fontconfig error: "/etc/fonts/conf.d/65-khmer.conf", line 23: out of
> > > memory
> > > Fontconfig error: "/etc/fonts/conf.d/65-khmer.conf", line 32: out of
> > > memory
> >
> > Are you really running out of memory?  If so, then that will cause you
> > lots of problems.  That may be why you are getting zombies.
>
> Hardly. It is a machine with 8 GB. I think I have seen the message for
> other software two and it was never a problem in practice. I think EVO
> always showed this, but it worked fine earlier.
>
> > > < HTTP/1.1 401 Unauthorized
> > > < Soup-Debug-Timestamp: 1392805687
> > > < Soup-Debug: ESoapMessage 1 (0x7fc097021dd0)
> > > < Set-Cookie: mail.liu.se=R408630383; path=/
> > > < Server: Microsoft-IIS/7.5
> > > < Set-Cookie: exchangecookie=...; expires=Thu, 19-Feb-2015 10:28:07 GMT;
> > > path=/; HttpOnly
> > > < WWW-Authenticate: ...
> > > < WWW-Authenticate: Negotiate
> > > < X-Powered-By: ASP.NET
> > > < Date: Wed, 19 Feb 2014 10:28:06 GMT
> > > < Content-Length: 0
> >
> > I don't think the 401 is a good sign ...
>
> It is the only 401. But I have no idea what the problem is here.
>
> > > > The cache is stored in ~/.cache/evolution/calendars - you could try
> > > > shutting down Evolution (and all the backend processes), and rename
> > > > that
> > > > folder - Evolution should recreate it all when you restart.
> > > How do I shut-down the backend processes? I google it, but could only
> > > find
> > > people wanting to remove EVO altogether. 'service' does not show
> > > evolution
> > > backends.
> >
> > Service only shows system things, not user land things.
> >
> > After you shutdown Evolution (do 'evolution --force-shutdown' if
> > necessary) you need to do
> >
> >  ps -ef | grep evolution
> >
> > to see any background Evolution processes that are still running (the
> > gnome calendar gets info from Evo and the alarm notify daemon is left
> > running).  Kill the processes using the 'kill' command.
>
> I tried, but the processes are relaunched all the time. I tried to move
> the cache, but the destination was removed automatically and now EVO
> crashes completely as soon as I switch to the calendar view. I will try
> once more as instructed below. If also that fails, I will try on a
> different account.

Experience with the new, empty account:
EVO freezes when switching to Calendar. I let it go and after some time
items from my account pop up.
I could change back to Email. However, when I switched back to Calendar
and pressed on work-week view, it freezes again.

I checked the process table, EVO is on about 43 minutes, while running for
about 3 hours. My machine has a Xeon E5-1620 (8x3.7GHz). EVO calendar
backend is at 7 minutes.

While writing: EVO is still frozen, more than 10 minutes now.

Any ideas?

As said before, some versions earlier, EVO was working quite OK with the
Exchange server. Just about one crash per day (~100 emails per day).

Michael

> Thanks
>
> Michael
>
> > If you really want to be sure, reboot your system, but don't login.
> > Switch to a console (Ctl-Alt-F2) and login there.  Make sure nothing Evo
> > related is running and then rename the cache folder.  Logout of the
> > console, switch back to the GUI (Alt-F1) and go from there.
> >
> > P.