[Fusionforge-general] mailfwd_update.pl hanging trying to get effective uid

[Sylvain Beucler - Inria]

Hi Seth,

(remember to "Reply to All" otherwise other people on the list don't see 
your answer)

mailfwd_update.pl doesn't seem critical, though make sure you're running 
this script as root.
(Note: I haven't worked with 5.2, only 5.1 and 5.3 ;))

Cheers!
Sylvain

Le 2014-09-04 16:34, Seth Galitzer a écrit :
> Neither was installed. I usually do use unscd on my hosts, I must have
> forgotten it on this one. Once I installed it, the script no longer
> hangs. However, I do get many errors like this when I run it:
> Cannot setegid(25351 25351): Operation not permitted at
> ./mailfwd_update.pl line 40.
>
> This error occurs for about 10% of my user accounts. The groups all seem
> to exist. Is it safe to ignore these errors, or is this an indication
> that my migration did not complete correctly?
>
> Thanks.
> Seth
>
> On 09/04/2014 03:05 AM, Sylvain Beucler - Inria wrote:
>> Hi Seth,
>>
>> First, can you make sure that nscd or unscd is installed?
>>
>> Cheers!
>> Sylvain
>>
>> Le 2014-09-03 21:50, Seth Galitzer a écrit :
>>> I'm finally going live with my FF5.2 host migrated from GFAS. I noticed
>>> today that when the update-user-group-ssh.sh script is run (either
>>> manually or from cron, the above-mentioned script hangs when trying to
>>> get the effective uid before writing out the new .forward file. The
>>> first couple of calls work fine, but it always hangs on the same user
>>> which is the third record returned.
>>>
>>> I've done a little debugging and the script hangs on the call to
>>> getpwnam() in GetUserUidGid() (line 63). I've tried calling that
>>> function on my own in a separate script on the same user acct that it
>>> hangs on and it seems to complete just fine and return the correct data.
>>> I also checked the sql statement that generates the list of users to
>>> process within psql and it doesn't seem to have any surprises.
>>>
>>> I've commented out the lines in the main script body that handle setting
>>> the effective uid for now, just so the cronjobs complete correctly, but
>>> I understand this could result in a race condition. My volume is low
>>> enough I'm hoping it won't be an issue for me.
>>>
>>> Any ideas on what might be happening here? I'm running v 5.2.3 on debian
>>> wheezy.
>>>
>>> Thanks.
>>> Seth
>>>