-->

DEVOPSZONES

  • Recent blogs

    SAN disk IO errors ("Sense Key : Illegal Request [current]")

    Error:

    SAN disk IO errors ("Sense Key : Illegal Request [current]")
    Oct 2 05:05:07 example.com kernel: sd 7:0:15:5: [sdaw]  Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
    Oct 2 05:05:07 example.com kernel: sd 7:0:15:5: [sdaw]  Sense Key : Illegal Request [current]
    Oct 2 05:05:07 example.com kernel: sd 7:0:15:5: [sdaw]  Add. Sense: Logical unit not supported
    Oct 2 05:05:07 example.com kernel: sd 7:0:15:5: [sdaw] CDB: Read(10): 28 00 00 00 00 00 00 00 08 00
    Oct 2 05:05:07 example.com kernel: end_request: I/O error, dev sdaw, sector 0
    ******************

    [root@example.com ~]# lvs
      /dev/sdq: read failed after 0 of 4096 at 0: Input/output error
      /dev/sdaw: read failed after 0 of 4096 at 0: Input/output error
      /dev/sdai: read failed after 0 of 4096 at 0: Input/output error
      /dev/sde: read failed after 0 of 4096 at 0: Input/output error
      /dev/sdx: read failed after 0 of 4096 at 0: Input/output error
      /dev/sdj: read failed after 0 of 4096 at 0: Input/output error
      /dev/sdap: read failed after 0 of 4096 at 0: Input/output error
      /dev/sdad: read failed after 0 of 4096 at 0: Input/output error
    ***********************
    Cause:

    These I/O errors refer to multiple causes in the system.  The Point Number 2 has more relevance to our problem.

    1. If SAN disks are in Replication Mode these are expected messages.
    2. There has been changes (Modify/Delete) to these LUNs outside of OS, which forced system to check the disks that are no longer present in the system.

    ***********************
    Diagnosis:

    1. Check if the disks are present in the system.
    fdisk -l
    2. If Yes, check if it has a Failed Path.
    multipath -ll
    3. Now Fire a Command "multipath -v0"  and check the Path status.
    4. If it has not come yet, do "service multipathd reload"
    5. If disk is not present in the system, then we need to clean the Kernel Entries to stop this errors.

    Resolution:

    1. Do rescan for all the device that are throwing error.

    echo 1 > /sys/class/scsi_device/7\:0\:3\:5/device/rescan

    2. Now the error should not be in the system.
      
      
      

    No comments