Re: [mlmmj] Diagnostic-Code: x-unix; internal software error / Command time limit exceeded: "/usr/lo

Sebastian Wolfgarten <[email protected]> Mon, 21 Mar 2016 21:07:34 +0000
Newsgroups	org.mlmmj.mlmmj
Message-ID	<[email protected]>
Dear Ben,

many thanks for following this up - I have largely increased the command_time_limit to 3000s, let’s see whether this fixes the issue. However I am confident it does.

Again, many thanks.

Best regards
Sebastian

> Am 20.03.2016 um 23:52 schrieb Ben Schmidt <[email protected]>:
> 
> Hi, Sebastian,
> 
> Yes, your config looks OK.
> 
> I think your answer might be here:
> 
> http://www.postfix.org/postconf.5.html#command_time_limit
> 
> A default of 1000 seconds would correspond roughly to Christian's report of "20 minutes later" as 1000 seconds is 16m40s.
> 
> Perhaps have a look with `postconf' to see if that value is still in place.
> 
> It can be overridden in a transport-specific way, too:
> 
> http://www.postfix.org/postconf.5.html#transport_time_limit
> 
> (I guess this means the directive would be mlmmj_time_limit in your case.)
> 
> Again, perhaps check with postconf.
> 
> I would expect you to see *something* in your mail.log about this. Perhaps grep it for "time limit exceeded" because I believe it might be just hidden at the end of a line about an attempted delivery, e.g. I found a similar report at <http://www-01.ibm.com/support/docview.wss?uid=swg21503353> where the log line in question was
> 
> Jun 20 09:39:18 postfix/pipe[11451]: 6ED022391325: to=<[email protected]>, relay=filter, delay=302, status=bounced (Command time limit exceeded: "/tmp/etc/postfix/wvpostqfilter")
> 
> It would probably be worth having a bit of a fish around to see if you can find those messages, though I doubt they give all that much additional information above what you already know.
> 
> Perhaps a next step would be to see if (in a dummy list and/or with a dummy transport) you can make it worse and get the same result if you decrease the configuration.
> 
> You may also want to ask the question as to whether your list is large enough and your system loaded enough that it warrants such a large delivery time. Perhaps something unrelated is going on on the server that is causing it to be slower than it should be and some general monitoring is in order.
> 
> Does that help?
> 
> Ben
> 
> 
> 
> On 17/03/2016 7:22 am, Sebastian Wolfgarten wrote:
>> Dear Ben,
>> 
>> thanks for listening and following up! I am happy to contribute to finding the root cause of this (potential) issue.
>> 
>> Here is a quick overview of what I have:
>> 
>> FreeBSD 10.1 (AMD64)
>> Postfix 3.1.0
>> mlmmj 1.2.17.1
>> 
>> So overall, pretty standard and nothing extraordinary I would say. Here is an overview of my config (basically taken from http://mlmmj.org/docs/readme-postfix/):
>> 
>> —
>> 
>> 1) Create mlmmi user
>> # id mlmmj
>> uid=1021(mlmmj) gid=1021(mlmmj) groups=1021(mlmmi)
>> 
>> 2) Create home dir and give it to mlmmi user
>> 
>> # ls -la /var/spool
>> (…)
>> drwxr-xr-x  33 mlmmj  mlmmj   45 21 Okt 22:38 mlmmj
>> 
>> 3) Amend configuration of Postfix's master.cf
>> 
>> # grep mlmmj /etc/postfix/master.cf
>> # mlmmj mailing lists
>> mlmmj   unix  -       n       n       -       -       pipe
>>    flags=DORhu user=mlmmj argv=/usr/local/bin/mlmmj-recieve -F -L /var/spool/mlmmj/$nexthop
>> 
>> 4) Amend main.cf of Postfix
>> 
>> # grep mlmmj /etc/postfix/main.cf
>> virtual_alias_maps = hash:/etc/postfix/virtual,hash:/var/spool/mlmmj/virtual,mysql:/etc/postfix/mysql_virtual_alias_maps.cf
>> transport_maps = regexp:/etc/postfix/transport,hash:/var/spool/mlmmj/transport
>> mlmmj_destination_recipient_limit = 1
>> recipient_delimiter = +
>> 
>> 5) Create file system structure for mailing list (flat file structure):
>> 
>> # ls -la /var/spool/mlmmj/newsletter
>> total 77
>> drwxr-xr-x  15 mlmmj  mlmmj    17  8 Nov 15:14 .
>> drwxr-xr-x  33 mlmmj  mlmmj    45 21 Okt 22:38 ..
>> drwxr-xr-x   2 mlmmj  mlmmj    14 11 Jan 18:06 archive
>> drwxr-xr-x   2 mlmmj  mlmmj   140  4 Mär 22:04 bounce
>> drwxr-xr-x   2 mlmmj  mlmmj     7 28 Sep 22:39 control
>> drwxr-xr-x   2 mlmmj  mlmmj     2  7 Sep  2015 digesters.d
>> drwxr-xr-x   2 mlmmj  mlmmj     2 10 Mär 21:26 incoming
>> -rw-r--r--   1 mlmmj  mlmmj     2  4 Mär 21:33 index
>> -rw-------   1 mlmmj  mlmmj  1234  4 Mär 21:33 mlmmj.operation.log
>> drwxr-xr-x   2 mlmmj  mlmmj    12 10 Mär 21:26 moderation
>> drwxr-xr-x   2 mlmmj  mlmmj     2  7 Sep  2015 nomailsubs.d
>> drwxr-xr-x   3 mlmmj  mlmmj     3 10 Mär 21:26 queue
>> drwxr-xr-x   2 mlmmj  mlmmj     2  7 Sep  2015 requeue
>> drwxr-xr-x   2 mlmmj  mlmmj     2  7 Sep  2015 subconf
>> drwxr-xr-x   2 mlmmj  mlmmj    27  7 Mär 06:13 subscribers.d
>> drwxr-xr-x   2 mlmmj  mlmmj    36  7 Sep  2015 text
>> drwxr-xr-x   2 mlmmj  mlmmj     2  7 Sep  2015 unsubconf
>> 
>> Note: In the operation.log there is nothing to suggest why the mailing list is malfunctioning sometimes.
>> 
>> 6) Creation/amendment of transport and virtual file for mlmmj and Postfix integration:
>> 
>> # grep newsletter /var/spool/mlmmj/transport
>> [email protected]   mlmmj:newsletter
>> 
>> # grep newsletter /var/spool/mlmmj/virtual
>> [email protected]    [email protected]
>> 
>> 7) Postmapping and then postfix restart
>> 
>> # postmap /var/spool/mlmmj/virtual
>> # postmap /var/spool/mlmmj/transport
>> # postfix reload
>> 
>> --
>> 
>> First question: I assume the config is correct, right?
>> 
>> Secondly, yes I checked all file permissions and they should be fine too.
>> 
>> Now the question is: How can I best troubleshoot the issue? Any further suggestions on where to look for a root cause? My mail logs did not really contain any further information.
>> 
>> Many thanks.
>> 
>> Best regards
>> Sebastian
>> 
>>> Am 14.03.2016 um 03:35 schrieb Ben Schmidt <[email protected]>:
>>> 
>>> I'm listening, I'm just a bit stumped.
>>> 
>>> I feel like if this was an Mlmmj bug, it would've manifested long ago. Nothing I know of has changed Mlmmj-wise, but this problem has just appeared in the last three months or so.
>>> 
>>> I wonder if something has changed recently in a relevant platform or MTA.
>>> 
>>> Could you confirm your Mlmmj, OS and MTA versions, and anything else you feel might be relevant about your environment?
>>> 
>>> It seems that something is enforcing a command time limit. Not sure if it's via the kernel or not, but I guess it's initiated if not enforced by the shell or MTA.
>>> 
>>> Ben
>>> 
>>> 
>>> 
>>> On 14/03/2016 6:45 am, [email protected] wrote:
>>>> (Resending message to see whether there is anyone who has seen this issue before
>>>> or can tell me how to troubleshoot this - thanks. And sorry for the double-posting.)
>>>> 
>>>> Hi,
>>>> 
>>>> I am running mlmmi 1.2.17.1 on FreeBSD 10 and I have come across a very strange
>>>> issue similar to the one described by a guy in December last year (see
>>>> http://mlmmj.org/archive/mlmmj/2015-12/0000000.html).
>>>> Basically I have a mailing list (besides many others that are working fine!) for
>>>> which some people receive emails, others are not getting any messages. The
>>>> moderators are sometimes also getting no messages to approve or they do receive
>>>> the following error:
>>>> —
>>>> This is the mail system at host waldfest.wolfgarten.com
>>>> <http://waldfest.wolfgarten.com>.
>>>> 
>>>> I'm sorry to have to inform you that your message could not
>>>> be delivered to one or more recipients. It's attached below.
>>>> 
>>>> For further assistance, please send mail to postmaster.
>>>> 
>>>> If you do so, please include this problem report. You can
>>>> delete your own text from the attached returned message.
>>>> 
>>>>                    The mail system
>>>> 
>>>> <elsaalumni.de--newsletter+moderate-7d4f69f9439871f4@localhost.mlmmj
>>>> <mailto:elsaalumni.de--newsletter+moderate-7d4f69f9439871f4@localhost.mlmmj>>: Command
>>>>     time limit exceeded: "/usr/local/bin/mlmmj-recieve"
>>>> 
>>>> --
>>>> Alternatively I have had situations, in which I am getting the „x-unix; internal
>>>> software error" message - anyone seen the same?
>>>> In the logs I cannot find anything which looks like an error - any ideas on how to
>>>> best troubleshoot this?
>>>> Thanks.
>>>> Best regards
>>>> Sebastian
>> 
> 
>
signature.asc (application/pgp-signature, 842 B)