[JBoss JIRA] (JGRP-2237) The single node in the cluster not become a coordinator after coordinator leave.
by Bela Ban (JIRA)
[ https://issues.jboss.org/browse/JGRP-2237?page=com.atlassian.jira.plugin.... ]
Bela Ban commented on JGRP-2237:
--------------------------------
Have you tried this with the config I attached (test.xml)? And please use the latest 4.0.8 to test; this is what I used and I never hit a problem.
Your log shows the member waited for 30s which is impossible as {{GMS.timeout}} is 3s in the config you attached! So either you used the wrong config, or attached the wrong config!
> The single node in the cluster not become a coordinator after coordinator leave.
> --------------------------------------------------------------------------------
>
> Key: JGRP-2237
> URL: https://issues.jboss.org/browse/JGRP-2237
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 4.0.2, 4.0.8
> Reporter: kfir avraham
> Assignee: Bela Ban
> Priority: Minor
> Attachments: test.xml
>
>
> I got cluster with 2 members, sometimes when the first node (coordinator) leave the cluster the second one is not become a coordinator.
> When the first one is rejoin, he could not determine coordinator and select new one from the nodes list.
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
8 years, 5 months
[JBoss JIRA] (JGRP-2238) KUBE_PING should read old env variables and should warn
by Bela Ban (JIRA)
[ https://issues.jboss.org/browse/JGRP-2238?page=com.atlassian.jira.plugin.... ]
Bela Ban updated JGRP-2238:
---------------------------
Fix Version/s: 4.0.9
> KUBE_PING should read old env variables and should warn
> -------------------------------------------------------
>
> Key: JGRP-2238
> URL: https://issues.jboss.org/browse/JGRP-2238
> Project: JGroups
> Issue Type: Enhancement
> Reporter: Galder Zamarreño
> Assignee: Galder Zamarreño
> Fix For: 4.0.9
>
>
> While I was upgrading Infinispan cluster on docker, I realised I was using old environment properties. These were neither read nor was any WARN messages being reported.
> The code should read the current properties and if not present should try to read the old ones and assign them. Then, we should check if both new and old properties are in use and we should also check if only old ones in use. Both scenarios should warn the user.
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
8 years, 5 months
[JBoss JIRA] (WFCORE-3435) Expose capability/requirements to deployments
by Tomaz Cerar (JIRA)
[ https://issues.jboss.org/browse/WFCORE-3435?page=com.atlassian.jira.plugi... ]
Tomaz Cerar commented on WFCORE-3435:
-------------------------------------
We should probably introduce something like CapabilityActivator.
Less breaking change would be having ServiceActivatorContext or implement some other interface that would also expose capabilities or have serviceTarget return CapabilityServiceTarget.
But both of this options would require that cast is done in ServiceActivator which could case other issues.
> Expose capability/requirements to deployments
> ---------------------------------------------
>
> Key: WFCORE-3435
> URL: https://issues.jboss.org/browse/WFCORE-3435
> Project: WildFly Core
> Issue Type: Enhancement
> Affects Versions: 3.0.9.Final
> Reporter: Bob McWhirter
> Assignee: Jason Greene
>
> For WildFly Swarm, we use a fair amount of ServiceActivators in deployments.
> Moving to WF11, we have seen some of our dependencies on Service<T>, such as NamingService.SERVICE_NAME result in deprecation warnings suggesting we move to using CAPABILITY_NAME. From within a ServiceActivator, within a deployment, this appears to not be a possibility.
> Filed per [~ctomc]
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
8 years, 5 months