[jira] [Resolved] (APLO-312) Apollo fails to bumb up the ulimit!

["Hiram Chirino (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/APLO-312?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Hiram Chirino resolved APLO-312.
--------------------------------

       Resolution: Fixed
    Fix Version/s: 1.7
         Assignee: Hiram Chirino

Ok, trunk just got improved ulimit bumping logic.  Not only can you configure the target ulimit by setting the APOLLO_ULIMIT shell var, but if setting it fails, it will do a binary search of the highest ulimit that it can set.

> Apollo fails to bumb up the ulimit!
> -----------------------------------
>
>                 Key: APLO-312
>                 URL: https://issues.apache.org/jira/browse/APLO-312
>             Project: ActiveMQ Apollo
>          Issue Type: Improvement
>          Components: apollo-broker
>            Reporter: Vaggelis Atlidakis
>            Assignee: Hiram Chirino
>             Fix For: 1.7
>
>
> Hi,
> I noticed that in the script: "/usr/share/apollo/bin/apollo" there is a section trying to maximize the limit of
> fds to be allocated by the broker. Unfortunately, this fails and the broker is limited to 4096 fds. It's no big 
> deal, but for our brokers we need a lot more active connections. 
> Can you consider adding a command like "ulimit -n 100000" in the init script? So that our puppet managed brokers have this limit automatically set?
> Best Regards,
> --
> aV



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)