Tuesday, 19 February 2019

How to rescan and recover LUN paths in a host after modifying SLM reporting nodes

Microsoft Windows hosts:
  • Rescan after add-reporting-nodes and remove-reporting-nodes using Windows GUI.
    1. Open Computer Management (Local)
    2. In the console tree, click Computer Management (Local) >> Storage >> Disk Management
    3. In the disk management page click Action >> Rescan Disks. This will rescan all the disks and update any path changes.
  • Rescan after add-reporting-nodes and remove-reporting-nodes using command line.
    1. Open Command Prompt and enter the following text:
      # diskpart
    2. At the DISKPART> prompt, enter the following text:
      DISKPART> rescan.
      This will rescan all the disks and updates any path changes. For more information, see Microsoft TechNet Updatedisk.
    3.  
      Linux hosts:
    4. Rescan after add-reporting-nodes.
      1. Starting RHEL 6.5 & RHEL 7.0 onwards, run the following command to update active/optimized paths afteradd-reporting-nodes:
        # /usr/bin/rescan-scsi-bus.sh –a
      2. For RHEL 5 and RHEL 6.4 (including previous updates), run the following command to update active/optimized paths after add-reporting-nodes:
        # /usr/bin/rescan-scsi-bus.sh
        Note: Nothing additional has to be done in the multipath layer.
    5. Rescan after remove-reporting-nodes
      1. Separate rescan steps are required for SCSI layer and Multipathing layer in Linux storage stack to clean up stale disk paths after remove-reporting-nodes in SLM.
      2. Run the following command to remove stale LUN paths in SCSI layer
        # /usr/bin/rescan-scsi-bus.sh –r
      3. Next run the following command to remove stale LUN paths in multipath layer:
        # multipath -r
        Solaris hosts:
      4. Rescan after add-reporting-nodes
        1. For iSCSI LUNs, run the following command:
          # devfsadm -i iscsi
        2. For FC/FCoE LUNs, perform the following steps:
          1. Run the following command to identify OS Device name of the HBA ports that are accessing NetApp LUNs:
            # cfgadm -al -o show_FCP_dev | grep fc-fabric
            c3 fc-fabric connected configured unknown
            c4 fc-fabric connected configured unknown
          2. Now run the following command for each <controller> to be rescanned:
            # cfgadm -c configure <controller>
            For example from Step1 c3 & c4 are the controller names and so the command would be:
            # cfgadm -c configure c3
            # cfgadm -c configure c4
      5. Rescan after remove-reporting-nodes
        1. For iSCSI LUNs, run the following command:
          # devfsadm -i iscsi
          # devfsadm -Cv
        2. For FC/FCoE LUNs, perform the following steps:
          1. If the host is accessing NetApp LUNs using a single FC port, then it is advised to reboot the host. Run the following commands to reconfigure and reboot the host.
            # touch /reconfigure
            # init 6
          2. But if host is accessing NetApp LUNs with 2 or more FC ports, then run the following commands to identify OS Device names of the HBA ports:
            # cfgadm -al -o show_FCP_dev | grep fc-fabric
            c3 fc-fabric connected configured unknown
            c4 fc-fabric connected configured unknown
          3. Run the following command to reconfigure each port one after the other:
            # cfgadm -c unconfigure <controller>
            # cfgadm -c configure <controller>

            For example from above output c3 & c4 are the controller names and so the commands would be similar to the following:
            # cfgadm -c unconfigure c3
            # cfgadm -c configure c3
            # cfgadm -c unconfigure c4
            # cfgadm -c configure c4

            Note: Above step should be peformed only for one port at a time.
          4. Run the following command to clean up the devices:
            # devfsadm -Cv
          5. To clear MPxIO entries, an OS reboot is needed and this can be performed during a planned downtime. Run the following command to reconfigure and reboot the host:
            # touch /reconfigure
            # init 6
          6. Once the host is back after reboot, run the following command :
            # devfsadm -Cv
      6.  

No comments:

Post a Comment