[rt-users] problems with mailgate

[jamie baddeley]

Hi Ruslan,

disk space is not a problem.

crm:/# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/hda3             1.9G  282M  1.6G  16% /
tmpfs                 110M     0  110M   0% /dev/shm
/dev/hda1              93M  5.0M   84M   6% /boot
/dev/hda4             3.5G  304M  3.1G   9% /var
crm:/#

I'm running rt3.0.10 on debian (unstable) - so it's a pretty new
version...

I noticed other posts on the list talking about similar problems, and I
noticed that recently this appeared in my logs:

 Couldn't write attachments to temp dir on disk. using more memory and
processor. (/usr/share/request-tracker3/lib/RT/Template_Overlay.pm:330)

I don't know if the two are related, but something smells fishy. - BTW
This could be a diversion, but it seems that since I've used the rt-cli,
the mailgate hasn't worked since....

jamie


On Fri, 2004-05-28 at 23:56, Ruslan U. Zakirov wrote:
> Jamie, RT has problems in old versions with cleanning tmp files, may be 
> your tmp partition run out of free space?
> 
> 
> jamie.baddeley at fx.net.nz wrote:
> > Hi,
> > 
> > I have had rt pretty running happily for a couple of months now. Then,
> > today, mailgate stopped working, with exim complaining:
> > 
> > 2004-05-28 20:00:01 1BTcHO-0000Ho-00 ** |/usr/bin/rt-mailgate --queue
> > General --action correspond --debug --url http://crm.fx.net.nz
> > <general at crm.fx.net.nz> D=rt_aliases T=address_pipe: Child process of
> > address_pipe transport (running command "/usr/bin/rt-mailgate --queue
> > General --action correspond --debug --url http://crm.fx.net.nz") was
> > terminated by signal 9 (Killed)
> > 
> > Exim spat this back at me - It said (via email):
> > 
> > RT server error. The RT server which handled your email did not behave
> > as expected. It said: 
> > 
> > System error
> > 
> > error:
> > Error in tempdir() using XXXXXXXXXX: Parent directory (.) is not
> > writable at /usr/share/request-tracker3/lib/RT/EmailParser.pm line 595
> > 
> > context:
> > ...
> > 82:     # whether they should generate a full stack trace (confess() and
> > cluck())
> > 83:     # or simply report the caller's package (croak() and carp()),
> > respectively.
> > 84:     # confess() and croak() die, carp() and cluck() warn.
> > 85:
> > 86:     sub croak { die shortmess @_ }
> > 87:     sub confess { die longmess @_ }
> > 88:     sub carp { warn shortmess @_ }
> > 89:     sub cluck { warn longmess @_ }
> > 90:
> > ...
> > code stack:     /usr/share/perl/5.8/Carp.pm:86
> > /usr/share/perl/5.8/File/Temp.pm:1360
> > /usr/share/request-tracker3/lib/RT/EmailParser.pm:595
> > /usr/share/request-tracker3/lib/RT/EmailParser.pm:233
> > /usr/share/request-tracker3/lib/RT/EmailParser.pm:177
> > /usr/share/request-tracker3/lib/RT/Interface/Email.pm:443
> > /usr/share/request-tracker3/html/REST/1.0/NoAuth/mail-gateway:31
> > 
> > I'm flummoxed. I didn't do anything AFAIK. 
> > 
> > Has anyone seen anything like this - did you find a resolution?
> > 
> > jamie
> >