Why does a cluster fail to activate a random VG/LV using CLVM when one of the nodes rejoins the cluster?
Issue
- Activating a large number of VGs fails on a random one when one of the nodes rejoins the cluster
- On the node that is trying to activate the VGs we see errors like the one below:
rgmanager[xxxx]: [lvm] Unable to perform required deactivation of cluster_vg/random_lv before starting
Environment
- Red Hat Enterprise Linux 6 with Resilient storage Add-On
rgmanager
- CLVMD
- A large number of clustered VGs (through clvmd) that are managed by
rgmanager
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.