Previous Topic: Remove Server with the BFC CLINext Topic: Removing Junction Points

Reference Information › CA AppLogic Support Knowledge Base › Overview of Support Knowledge Base › Remove RAID from Running Grid / Removing Nodes for Maintenance AppLogic 3.x

Remove RAID from Running Grid / Removing Nodes for Maintenance AppLogic 3.x

Partner and Customer created initial test grid (5 nodes) with a RAID 5 and then CA services added the rest of the 5 nodes to the grid without any RAID enabled. Our goal is to disable the RAID 5 and re-add these nodes to the grid.

Current configuration: There are 10 total test grid nodes, 5 of which have RAID 5 enabled.

  1. ssh to the controller
  2. Check controller volumes with: 
    vol info _SYSTEM:boot
vol info _SYSTEM:meta
vol info _SYSTEM:impex

    Our goal is to verify that these volumes are hosted on nodes other than those which are a part of the RAID and that they have “synced=1” which means that the volume is in a good state on that node. If the controller is hosted on the node you may want this to be the last server to be migrated as the controller will failover to the secondary node when it is brought down.

  3. 3t vol list --all |grep srvX - where srvX is the server which you want to bring down to remove the RAID. Perform this to verify that all volumes which are stored on this node also have a good mirror on another node. If volumes do not have another mirror, they must be migrated manually with the ‘3t vol migrate <volume>’ command. Example: 3t vol migrate /user:ISO2CLASS.boot If you receive the message “does not need to be migrated.” then it should still be successfully migrated in step 8.
  4. Login to the BFC UI and quarantine the node in which you will be working with. If this does not work, try dropping the number of ‘Target’ grid servers down a notch and try to quarantine / force removal. If these steps do not work go ahead and continue.
  5. 3t srv disable srvX - srvX is the server which you want to bring down to remove the RAID.
  6. 3t vol migrate –all - ensures all volumes which were on the node which is now disabled are migrated to the rest of the grid and are in good standing. I recommend running this in a Screen session. You may view the status by opening up another window or disconnecting from the screen session and running ‘3t vol migrate –status’
  7. Once the migrate is complete you may reboot the server into the HP tools disk to change the RAID controller settings, ensuring the RAID is disabled.
  8. Reboot the node once the RAID settings have been updated. The node should PXE boot.
  9. Unquarentine the node from the BFC UI. If you dropped the ‘Target’ number of grid servers you'll need to raise it up a notch to allow this node to be re-added to the grid.
  10. 10. If the BFC does not pick up the node (not recommended to pass along to the customer).
    1. A. From the BFC terminal, cd to the installation directory which should be ‘cd /opt/bfc/applogic-versions/3.1.14’
    2. B. ./aldo remsrv srvX <grid_Name>
    3. C. Reboot the node and it should then be rediscovered by the BFC
    4. D. Raise the ‘Target’ number of grid servers up a notch if necessary to allow this node to be re-added to the grid.

    Continue these steps until the RAID has been removed from all necessary nodes.