The mailing list for listmasters using Sympa

["Silvers, Tim" <address@concealed>]

Hi Matt,

 

Our graphs are not as detailed, but they are attached. I provided the graph from this week when we experienced a crash and a graph covering the year (our upgrade from 6.1.x was completed in May). Today, I added more details to our graphs, so you’ll see that reappear in stats from today in the weekly.

 

Thanks,

 

Tim

 

 

From: address@concealed <address@concealed> on behalf of Matt Taggart <address@concealed>
Date: Wednesday, December 8, 2021 at 4:53 PM
To: address@concealed <address@concealed>
Subject: [External] Re: [sympa-users] Sympa process died

This message was sent from a non-IU address. Please exercise caution when clicking links or opening attachments from external sources.
-------

On 12/8/21 1:37 PM, Silvers, Tim wrote:
> We have recently started experiencing issues with Sympa crashing. Logs
> will indicate memory cannot be allocated. Memory usage on the system
> during the crashes will often be 85-90% used.
>
> Here is the log line from the latest crash:
>
> Dec  7 17:25:09 list-app sympa_msg[1787]: err main::#243 >
> Sympa::Spindle::spin#95 > Sympa::Spindle::DoForward::_twist#186 >
> Sympa::Mailer::store#160 > Sympa::Mailer::_safefork#255 DIED: Exiting
> because cannot create new process for
> <address@concealed>: Cannot allocate memory at
> /usr/share/sympa/lib/Sympa/Mailer.pm line 255.
>
> Line 255 in Mailer.pm is the die statement inside of the ‘safefork’
> subroutine [1].
>
> We are running Sympa 6.2.64 on CentOS 7 and we have 8 GB of RAM
> allocated to the system.

Do you have any graphs of memory usage? Like grafana or munin?
Our sympa install uses quite a bit all the time, plus there are all the
system caches, etc. I suspect 8gb might be limiting you, but a graph
would probably show for sure.

Attached an example of ours.

--
Matt Taggart
address@concealed

Attachment: Sympa RAM Week.png
Description: Sympa RAM Week.png

Attachment: Sympa RAM Year.png
Description: Sympa RAM Year.png