reservation conflicts in a RHEL 7 High Availability cluster using fence_scsi after adding a new storage device to the cluster
Issue
- I added a new device to the
devices
list on myfence_scsi
stonith
device and now I'm getting reservation conflicts in the logs and I/O errors when accessing that device. - After adding a new device to the working
fence_scsi
stonith device the device fails to register properly.
Apr 23 11:14:40 rhel7-node2 kernel: [69591.038343] sd 2:0:0:0: reservation conflict
Apr 23 11:14:40 rhel7-node2 kernel: [69591.042034] sd 2:0:0:0: [sdd] Unhandled error code
Apr 23 11:14:40 rhel7-node2 kernel: sd 2:0:0:0: reservation conflict
Apr 23 11:14:40 rhel7-node2 kernel: sd 2:0:0:0: [sdd] Unhandled error code
Apr 23 11:14:40 rhel7-node2 kernel: sd 2:0:0:0: [sdd]
Apr 23 11:14:40 rhel7-node2 kernel: [69591.048040] sd 2:0:0:0: [sdd]
Apr 23 11:14:40 rhel7-node2 kernel: [69591.054034] Result: hostbyte=DID_OK driverbyte=DRIVER_OK
Apr 23 11:14:40 rhel7-node2 kernel: Result: hostbyte=DID_OK driverbyte=DRIVER_OK
Apr 23 11:14:40 rhel7-node2 kernel: [69591.060041] sd 2:0:0:0: [sdd] CDB:
Apr 23 11:14:40 rhel7-node2 kernel: sd 2:0:0:0: [sdd] CDB:
Apr 23 11:14:40 rhel7-node2 kernel: [69591.063062] Write(10): 2a 00 00 00 00 48 00 00 10 00
Apr 23 11:14:40 rhel7-node2 kernel: [69591.072049] blk_update_request: 263 callbacks suppressed
Apr 23 11:14:40 rhel7-node2 kernel: Write(10): 2a 00 00 00 00 48 00 00 10 00
Apr 23 11:14:40 rhel7-node2 kernel: blk_update_request: 263 callbacks suppressed
Apr 23 11:14:40 rhel7-node2 kernel: [69591.078040] end_request: critical nexus error, dev sdd, sector 72
Apr 23 11:14:40 rhel7-node2 kernel: end_request: critical nexus error, dev sdd, sector 72
Apr 23 11:14:40 rhel7-node2 kernel: [69591.087043] device-mapper: multipath: Failing path 8:48.
Apr 23 11:14:40 rhel7-node2 kernel: device-mapper: multipath: Failing path 8:48.
- We added a new storage device to our cluster, but
fence_scsi
hasn't unfenced it. - Why hasn't the cluster registered the node keys to a new device that we added when using
fence_scsi
?
Environment
- Red Hat Enterprise Linux (RHEL) 7 with the High Availability Add On
fence-agents-scsi-4.0.11-11.el7
or later- One or more
stonith
devices configured to usefence_scsi
as the agent- Device is configured with `meta provides="unfencing". If this is not configured, it could be an alternate explanation for the lack of unfencing.
- A new storage device has been added to the
device
list or to a clustered volume group after some had already existed and been "unfenced"
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.