Description of problem: ----------------------- There were 3 nodes that runs virt + gluster service in RHHI-V environment, when the environment is upgraded with in-service fashion, on one of the node, after reboot, glusterd holds incorrect information regarding the brick port gluster volume status mentions that the brick is listening to port 49154, but the process is actually listening to 49152 Version-Release number of selected component (if applicable): -------------------------------------------------------------- RHGS 3.5.0 How reproducible: ----------------- Only once Steps to Reproduce: -------------------- 1. Created a node with glusterfs-6.0-7 (interim build ) 2. Upgraded one of the node in a in-service upgrade fashion, and rebooted the node. Upgraded gluster to 6.0-9 3. Post reboot, checked for output of 'gluster volume heal' Actual results: --------------- 'gluster volume status' indicates that the brick is using the port 49154, but in actual the brick is using the port 49152 Expected results: ------------------ glusterd should show consistent result with actual process.
Created attachment 1601886 [details] sosreport-from-grafton10-node
Created attachment 1601888 [details] glusterd.log_from_grafton10