I have a cluster with two nodes. I have checked on both nodes if the disk is supporting;
[root@d-clus1 ~]# /usr/bin/sg_persist -n -i -r -d /dev/sdb
PR generation=0x1, Reservation follows:
Key=0xec290000
scope: LU_SCOPE, type: Write Exclusive, registrants only
[root@d-clus2 ~]# /usr/bin/sg_persist -n -i -r -d /dev/sdb
PR generation=0x1, Reservation follows:
Key=0xec290001
scope: LU_SCOPE, type: Write Exclusive, registrants only
I have configured the scsi stonith;
[root@d-clus1 ~]# pcs stonith create scsi fence_scsi pcmk_host_list="d-clus1 d-clus2" pcmk_reboot_action="off" pcmk_monitor_action="metadata" devices="/dev/sdb" meta provides="unfencing" --force
I is successfully started, however both node is reading only its own key;
[root@d-clus1 ~]# /usr/bin/sg_persist -n -i -k -d /dev/sdb
PR generation=0x1, 1 registered reservation key follows:
0xec290000
[root@d-clus2 ~]# /usr/bin/sg_persist -n -i -k -d /dev/sdb
PR generation=0x1, 1 registered reservation key follows:
0xec290001
And it is causing the failure on fencing. Where is my mistake?