Access to a filesystem in IBM cluster fails with 'reservation conflict' errors

Issue

• Access to a filesystem in IBM cluster fails with reservation conflict errors.

• We cannot mount xfs file systems after a SAN outage on IBM cluster node. Below are the errors observed while trying to mount this filesystem:

  :::GBLRESRM_MONITOR_FAIL#012IBM.Application could not run the monitor command.#012Resource name#012instancehost_pdbxxx_cluster-node1
  Oct 30 02:30:55 cluster-node1 GblResRM[7706]: (Recorded using libct_ffdc.a cv 2):::Error ID: :::Reference ID:  :::Template ID: 0:::Details File:  :::Location: RSCT,Application.C,
  :::GBLRESRM_MONITOR_FAIL#012IBM.Application could not run the monitor command.#012Resource name#012primary_pdbxxx_900-rs
  [...]
  Oct 30 02:30:58 cluster-node1 kernel: sd 4:0:1:1: reservation conflict
  Oct 30 02:30:58 cluster-node1 kernel: sd 4:0:1:1: [sdav] tag#2 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK
  Oct 30 02:30:58 cluster-node1 kernel: sd 4:0:1:1: [sdav] tag#2 CDB: Write(10) 2a 00 00 a0 36 d4 00 02 00 00
  Oct 30 02:30:58 cluster-node1 kernel: blk_update_request: critical nexus error, dev sdav, sector 10499796
  Oct 30 02:30:58 cluster-node1 kernel: sd 4:0:0:1: reservation conflict
  Oct 30 02:30:58 cluster-node1 kernel: sd 4:0:0:1: [sdb] tag#1 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK
  Oct 30 02:30:58 cluster-node1 kernel: sd 4:0:0:1: [sdb] tag#1 CDB: Write(10) 2a 00 00 a0 34 d4 00 02 00 00
  [...]