Description of problem: Sometimes, clustat starts dumping ominous messages to the log file. This is caused by clustat having the wrong authentication data after a total cluster outage (ie, "service cluster stop" on both nodes). Other side effects are listed below. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Run clustat. 2. Stop the cluster services on both nodes. 3. Restart the cluster services on both nodes. Actual Results: - The above log messages - "output of memcmp: -1" from cluquorumd and clupowerd - clustat showing "Unknown error 4294967295" for power switch status on one or both nodes. Expected Results: Unspecified at this time. Additional info: Restart clustat after a total cluster outage. Considering making clustat restart itself after it detects a cluster outage, or at least, reread the cluster authentication data.
Fix for clustat in pool.