Red Hat Bugzilla – Bug 1287094
A-MQ hawt.io console deleting queue does not remove it from the tree view
Last modified: 2016-09-07 17:03:18 EDT
Description of problem:
After deleting a Queue in the ActiveMQ Hawt.io console, the queue is still visible in the tree view on the left.
Opening the console again make the queue disappear, so this is really just the view not being refreshed.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Update xPaaS A-MQ image stream to use rcm-img-docker01.build.eng.bos.redhat.com:5001/jboss-amq-6/amq62-openshift:1.2-6 (as the previous version lack the jolokia integration for the hawt.io console to work)
2. Deploy A-MQ using the updated xPaaS template
3. locate the A-MQ pod in the OpenShift console and click 'Connect'
4. Crate a 'fooBar' queue.
5. Delete the 'fooBar' queue
Queue is deleted, but is still visible in the Queue folder in the tree view on the left
Removed queue should disappear also from the treee view.
Reproducible with both Firefox and Chrome
It should be '5000', and actually is supposed to be configurable but I see that particular preference page isn't available, so will need to fix that for starters.
Not a fix for the specific issue, but since I mentioned I'd fix it, this commit adds back in the update rate setting into the preferences, will show up in the 'Jolokia' page:
Still waiting to hear back, as I can't seem to reproduce this locally at least.
localStorage['updateRate'] is 5000
I see stream of POST requests to https://cloudenablement-qe4.lab.eng.brq.redhat.com:8443/api/v1/namespaces/maschmid/pods/https:amq-1-9jfsr:8778/proxy/jolokia/?maxDepth=7&maxCollectionSize=500&ignoreErrors=true&canonicalNaming=false once per 5 seconds.
The removed queue will disappear if I switch to "JMX" or "Threads" view and back to "ActiveMQ" again, so at least that seem to refresh the tree.
Fix mentioned above is in this PR -> https://github.com/openshift/origin/pull/6539
fix is not included in current latest puddle, will verify once merged in new OSE puddle
Still reproducible with the current puddle (openshift-jvm Version: 1.0.39 Commit ID: e79966878b9aba11b77fcb69db2cbbba12c662c9 ), oc v184.108.40.206, kubernetes v1.1.0-origin-1107-g4c8e6f4 )
Used the latest A-MQ image from RCM ( rcm-img-docker01.build.eng.bos.redhat.com:5001/jboss-amq-6/amq62-openshift:1.2-12 )
Realized that this is by design in a way. We avoid fetching the JMX tree regularly as it can contain a lot of data. Normally we get around this by keeping an eye on a special TreeWatcher mbean, however if this mbean isn't available in the tree we have to degrade to refreshing the tree less.
Will probably look at adding a tree refresh when the delete operation completes.
Here's the fix -> https://github.com/hawtio/hawtio-integration/commit/d5bd0fc642e12dcf6be14bacdaef8d786cd1b278
Will be in this openshift-jvm version -> https://github.com/hawtio/openshift-jvm/commit/d46658f96d3134f99b0b6073433d4366defaee0f
Will update once I've submitted a PR...
PR => https://github.com/openshift/origin/pull/6938
Verified on current 3.2 Puddle (the tree is still not updated in another browser window when queues are added or deleted, but as you said, that is by design to not refresh the tree)
(openshift-jvm Version: 1.0.42 Commit ID: d46658f96d3134f99b0b6073433d4366defaee0f )