You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@activemq.apache.org by "Michele Bozzaotre (JIRA)" <ji...@apache.org> on 2017/02/20 18:27:44 UTC

[jira] [Comment Edited] (AMQCPP-601) PriorityBackup isn't working

    [ https://issues.apache.org/jira/browse/AMQCPP-601?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15874917#comment-15874917 ] 

Michele Bozzaotre edited comment on AMQCPP-601 at 2/20/17 6:27 PM:
-------------------------------------------------------------------

Hi Tim,
Thanks for your fix! It seems to work very well now, except for the case I'm going to describe below.
Same scenario of the previous comments: a client connected to a cluter with two nodes by the URI:

failover:(tcp://local:61616,tcp://remote:61616)?randomize=false&priorityBackup=true

When one of the two hosts is unplugged from the network, the client is always able to switch to the other host very quickly, as expected in case of priority backup active.
However suppose to unplug local host and, just after the client has switched on remote host, try to restore local host connection. Now the client enters a loop where it switches very fast the connection between the two hosts, with a frequency of about half a second. During this swing, the client is still able to receive all the messages. Once you unplug one of the hosts, the client stops to switch and stayes connected to the other one.

Tim, let me know if you need more details or if Ican help you in some way.

Thanks


was (Author: michele.bozzaotre@ansaldo-sts.com):
Hi Tim,
Thanks for your fix! It seems to work very well now, except for the case I'm going to describe below.
Same scenario of the previous comments: a client connected to a cluter with two nodes by the URI:

failover:(tcp://local:61616,tcp://remote:61616)?randomize=false&priorityBackup=true

When one of the two hosts is unplugged from the network, the client is always able to switch to the other host very quickly, as expected in case of priority backup active.
However suppose to unplug local host and, just after the client has switched on remote host, try to restore local host connection. Now the client enters a loop where it switches very fast the connection between the two hosts, with a frequency of about half a second. During this swing, the client is still able to receive all the messages. Once you unplug one of the hosts, the client stops to switch and stayed connected to the other one.

Tim, let me know if you need more details or if Ican help you in some way.

Thanks

> PriorityBackup isn't working
> ----------------------------
>
>                 Key: AMQCPP-601
>                 URL: https://issues.apache.org/jira/browse/AMQCPP-601
>             Project: ActiveMQ C++ Client
>          Issue Type: Bug
>    Affects Versions: 3.7.0, 3.9.3
>         Environment: MSVC (Toolset v90), 
>            Reporter: Jawad Bokhari
>            Assignee: Timothy Bish
>             Fix For: 3.9.4
>
>
> I am trying to take advantage of the ActiveMQ priority backup feature.
> I'm running 2 brokers on 2 different sites (machines). Each site has 2 AMQ clients, one of which is a Java client and the other one is a C++ client.
> Both brokers are linked as network of brokers in duplex mode. 
> I want the following behavior for my apps.
> * Always connect to local broker on startup
> * If local broker goes down, connect to the remote
> * {color:red} While connected to remote, if local comes back up, we then reconnect to local. {color}
> This works on my java app by simply adding priorityBackup to my URI options 
> ??failover:(tcp://local:61616,tcp://remote:61616)?randomize=false&priorityBackup=true??
> However, the part highlighted in {color:red}red{color} doesn't work for CPP client.
> The following works fine on the CPP apps (with basic working failover functionality - aka jumping to remote when local goes down )
> ??failover:(tcp://local:61616,tcp://remote:61616)?randomize=false??
> But updating the uri options with priorityBackup even breaks failover functionality completely (my apps never failover to the remote broker, they just stay in some kind of broker-less/limbo state when their local broker goes down) 
> ??failover:(tcp://local:61616,tcp://remote:61616)?randomize=false{color:red}&priorityBackup=true{color}??
> I tried this with 3.9.3 without success. Then, I tried with the version 3.7.0 when this feature *priorityBackup* was first introduced but without luck.
> ActiveMQ broker versions: *5.9.0*, *5.13.2*
> I'm using Visual Studio 2008 toolset (V90) for my client application. 



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)