LVM commands hang in cluster after clvmd is killed by oomkiller on Red Hat Enterprise Linux 5
Issue
- Our system ran out of memory and oomkiller killed clvmd. Afterwards, LVM commands were hanging and our LVM mirrored LUN's were operating slowly.
Environment
- Red Hat Enterprise Linux (RHEL) including:
- Red Hat Enterprise Linux Server 5 (with the High Availability and Resilient Storage Add Ons)
- Issue was observed on RHEL5.7 with kernel 2.6.18-274.18.1.el5xen, but it is currently unknown if other versions are also affected.
- Red Hat High Availability cluster with 2 or more nodes
- Clustered LVM (with clvmd running). Issue has been observed with lvm2-2.02.84-6.el5_7.2, but it is currently unknown if other versions are also affected.
- cmirror is in use to mirror clustered LVM volumes.
- System ran out of memory and oomkiller killed clvmd:
kernel: xend invoked oom-killer: gfp_mask=0x200d2, order=0, oomkilladj=0
kernel: Out of memory: Killed process 24949, UID 0, (clvmd).
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
