Description of problem: ----------------------- Consider a cluster having a single volume, being monitored be Nagios. When the volume is utilized beyond the critical level, the volume utilization service reports the volume to be critical. The "host" that represents the cluster in the Nagios UI, reports "Cluster Status OK : None of the Volumes are in Critical State" as part of its status information. Version-Release number of selected component (if applicable): gluster-nagios-addons-0.1.0-25.git25f0bba.el6.x86_64 How reproducible: Saw it once. Steps to Reproduce: 1. Create a cluster of RHS nodes, and create a volume, start it and fill the volume with data, such that it crosses 90% capacity. 2. Configure nagios server to run on one of the RHS nodes, and configure this cluster to be monitored by nagios. 3. Check the volume utilization service for the volume created, it should show critical status. 4. Check the state information for the "host" that represents the cluster in the Nagios UI. Actual results: The status information for the host says "Cluster Status OK : None of the Volumes are in Critical State". Expected results: The status information should indicate that one of the volumes in the cluster is utilized above critical level. Additional info:
The performance data for the cluster also shows the number of volumes in critical state as zero - noOfVolumesInCriticalState=0
As per the current design, cluster status is determined from the status of 'Volume Status' services of all the volumes under that cluster. The 'Cluster Utilization' service will reflect if any of volume's utilization goes beyond critical level. When the cascading gets implemented(in future) the cluster utilization service's will get propagated to cluster status This is inline with the implementation of volume Status as well. Similarly we don't consider the Brick Utilization while determining the Volume status. So as per my opinion this is not a bug and works as per the design
As discussed in the triage meeting : we will take it up in the 3.1 release
Please review and signoff edited doc text.
Doctext is fine
What would be the expected behavior? - Cluster status will be aggregation of status of all volumes or - Cluster status will be aggregation of both status and utilization of volumes
Thank you for your report. However, this bug is being closed as it's logged against gluster-nagios monitoring for which no further new development is being undertaken.
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days