Fabric MetroCluster - scsi.cmd.abortedByHost:error
Applies to
Fabric MetroCluster
Issue
The following error messages are found in the event log:
Cluster1::> event log show -message-name scsi.cmd.abortedByHost
Time Node Severity Event
------------------- ------- ------------- ---------------------------
10/4/2020 07:09:46 Node01 ERROR Disk device brocade-fab1-sw1:13.126L1506: Command aborted by host adapter: ...
10/4/2020 05:16:42 Node01 ERROR Disk device brocade-fab1-sw3:13.126L542: Command aborted by host adapter: ...
10/4/2020 03:18:07 Node01 ERROR Disk device brocade-fab1-sw3:13.126L1: Command aborted by host adapter: ...
The above scsi errors are usually indicative of link / physical layer issues. Whenever there are link-layer issues there are also the
scsi.path.excessiveErrors:error
reported:Cluster1::> event log show -message-name scsi.path.excessiveErrors
Time Node Severity Event
------------------- ------ -------- ---------------------------
10/4/2020 05:42:46 Node01 ERROR Excessive errors encountered by adapter 11a on disk device brocade-fab1-sw1:13.126.
10/4/2020 05:16:42 Node01 ERROR Excessive errors encountered by adapter 11a on disk device brocade-fab1-sw3:13.126.
scsi.cmd.abortedByHost
reported for all disks on both local and remote switchesscsi.path.excessiveErrors
reported by a single port on a single node
Hence one needs to focus on the link between the port and the switch it is connected to. After statsclear
on the switch new CRC errors are being reported on the port where the initiator is connected:
brocade-fab1-sw1:admin> porterrshow
frames enc crc crc too too bad enc disc link
tx rx in err g_eof shrt long eof out c3 fail
29: 8.4g 2.4g 0 128k 118k 184L 0 10k 0 565k 113L