on 12-16-2016 3:31 PM
We are using Hybris 6.1
We have 5 nodes in the cluster. 3 are for storefront and 2 are used for the back office, cockpits etc. We are facing a issue where in cluster nodes [0-4] are displayed randomly and not all of them together. Sometimes it node 0 shows 0,2,4 but not 1,3, sometimes totally different cluster. Also this varies from cluster to cluster. But none of them shows all of them together.
We have verified the mcast udp details cluster configurations and i am running out of ideas.
Any pointers would be helpful.
below are the local.properties values pertaining to cluster we have #Cluster clustermode=true cluster.id=4 cluster.broadcast.methods=jgroups cluster.broadcast.method.jgroups=de.hybris.platform.cluster.jgroups.JGroupsBroadcastMethod cluster.broadcast.method.jgroups.channel.name=hybris-prd-broadcast cluster.broadcast.method.jgroups.configuration=jgroups-udp.xml cluster.nodes.autodiscovery=false cluster.nodes.ping.interval=10000 cluster.nodes.stale.timout=30000
task.processing.enabled = true for the 2 nodes (node id 3,4) which will be used for the back office, cockpits
task.processing.enabled = false for the nodes(node id 0,1,2) for storefront.
Appreciate your help.
We've experienced issues with JGroups binding to the wrong interface and/or sending to IPv6 addresses.
Add the following either to your tomcat.generaloptions
or tomcat.javaoptions
in local.properties
Don't forget you need to at least run ant deploy
after changing these properties to get them copied across to the runtime settings.
-Djava.net.preferIPv4Stack=true -Djava.net.preferIPv6Addresses=false
You may also find that binding to a specific interface helps too
-Djgroups.bind_addr=match-interface:eth0
obviously you need to choose the correct interface for your machine if you've multiple network adapters in it.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Check to see if any firewall may drop udp or tcp packets between hybris nodes, disable any firewalls between hybris nodes, restart hybris and try again
Check distribution of hybris nodes across hypervisor hosts, any patterns as it relates to nodes not talking to each other?
Use jgroups multicast sender and receiver utilities to test if nodes can talk to each other using those utilities, test both ways
If you have multiple network interfaces make sure you force hybris to bind on a consistent interface on all nodes
Make sure nodes can resolve each other by name consistently
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
6 | |
1 | |
1 | |
1 | |
1 | |
1 | |
1 | |
1 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.