This should work too without changing supervisor init script.

[program:redis-a]
command=bash -c "ulimit -n 32768; exec /usr/local/bin/redis-server
/etc/redis/a.conf"


A.

------------------------------------------------------
Ales Zoulek
+420 604 332 515
Jabber: [email protected]
------------------------------------------------------


On Fri, Dec 9, 2011 at 7:26 PM, Timothy Jones
<[email protected]>wrote:

>  The way I read the documentation on minfds was that it didn’t raise the
> limit, but only ensured that the limit was *at least*  that value before
> starting.****
>
> ** **
>
> In any case, I would suggest raising your ulimit –n value in the shell
> that starts supervisor.  Then the children of supervisor (all of them) will
> inherit the higher limit.  If you need a really high value (up to 65535),
> you will have to raise the hard limit in /etc/security/limits.d first.****
>
> ** **
>
> ** **
>
> tlj****
>
> ** **
>
> *From:* [email protected] [mailto:
> [email protected]] *On Behalf Of *Kaelten
> *Sent:* Friday, December 09, 2011 8:56 AM
> *To:* [email protected]
> *Subject:* [Supervisor-users] Clarification on ulimit and minfds****
>
> ** **
>
> I've done some reading on the configuration file and found a ticket that
> seemed to indicate that I can manipulate the file descriptors limit of a
> child process by setting the minfds in the supervisord config file.****
>
> ** **
>
> My situation is that I'm needing to run a redis instance that could allow
> for a lot of connections.  I run it directly from supervisor with no run
> script in between so I can't just call ulimit -n on it.****
>
> ** **
>
> Am I correct in understanding this?****
>
>
> Bryan McLemore
> Kaelten****
>
> _______________________________________________
> Supervisor-users mailing list
> [email protected]
> http://lists.supervisord.org/mailman/listinfo/supervisor-users
>
>
_______________________________________________
Supervisor-users mailing list
[email protected]
http://lists.supervisord.org/mailman/listinfo/supervisor-users

Reply via email to