[jira] [Commented] (AMQ-4497) Shared File System Master/Slave function broken with NFSv4 after network back stable

["Vitalii Samolovskikh (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AMQ-4497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14060559#comment-14060559 ] 

Vitalii Samolovskikh commented on AMQ-4497:
-------------------------------------------

A behavior of the ActiveMQ Master/Slave depends of the configuration of the NFS connection. For example if I mount a shared storage with the `soft` option and break the connection to the storage the master is crushed.

It is necessary to configure the NFS connection to the store thoroughly. But I can't give more recommendation about it, only use the `soft` option when mount a storage directory.

> Shared File System Master/Slave function broken with NFSv4 after network back stable
> ------------------------------------------------------------------------------------
>
>                 Key: AMQ-4497
>                 URL: https://issues.apache.org/jira/browse/AMQ-4497
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker
>    Affects Versions: 5.5.1
>         Environment: NFSv4
>            Reporter: Eden Chen
>
> In the experiment, I used kahaDB to be the persistenceAdapter. 
> The setting is like: 
> <persistenceAdapter>
> 　　<kahaDB directory="/nfs4/server/data/kahadb"/>
> </persistenceAdapter>
> There are 2 brokers in different machines, and the NFS server is installed in one of the machines.
>  
> I will use "broker1" to present the one without out NFS server and "broker2" for the other one. 
> I start broker1 earlier then brokerQ2, so the broker1 grab the file lock and be the master, broker2 be slave brokers and sit in a loop trying to grab the lock from the master broker. 
> Then I disconnect broker1's network, broker2 can grab the file lock and start working after a while, Broker1 is still alive and don't show any error message. 
> After broker2 starting working, I enable broker1's network, both broker1 and broker2 can serve client's request at the same time (I can use webconsole to send message for both brokers). 
> Both master and slave broker can grab the same file lock and receive client's request, this can cause losing message or restart AMQ failed. 



--
This message was sent by Atlassian JIRA
(v6.2#6252)