Description of problem: On an eight node cluster, we have configured 24 IP services. On most of the nodes, clustat times out and does not show the IP services. At most on two of the eight nodes clustat will show the IP services. Version-Release number of selected component (if applicable): rgmanager-1.9.46-0 How reproducible: Every time Steps to Reproduce: 1. Configure an 8 node cluster with 24 IP services 2. Start the cluster 3. Run clustat on each node Actual results: On at least 5 or 6 of the nodes clustat times out Expected results: Clustat to return complete status on all nodes Additional info:
Clustat also times out on 1 node of a 3 node cluster with 51 IP services configured. The test reproduction and environment info are exactly the same as reported by Henry with the exception being number of nodes and services.
I expect to have an initial pass at a fix by tomorrow morning. With some luck, I will have it today.
Sounds good, thanks.
Here's some additional info. We have found that if we boot all nodes with a smaller number of VIPs and then configure additional VIPs on a running cluster we can get more VIPs to run than if we boot all the nodes with the total number of VIPS already configured. Also, is the fix you are working on going to address the other bugs I've opened on rgmanager? It looks like they may all be related.
I believe this should fix this problem and 189777 problem... let me know: http://people.redhat.com/lhh/rgmanager-1.9.46-1speed.x86_64.rpm http://people.redhat.com/lhh/rgmanager-1.9.46-1speed.i386.rpm http://people.redhat.com/lhh/rgmanager-1.9.46-1speed.src.rpm
Thanks, Lon. We'll try it right away and let you know what happens.
*** This bug has been marked as a duplicate of 182454 ***