Deleting a node from a cluster using the CLI Advanced function cluster overview using the CLI Maintaining passwords using the CLI

This task provides step-by-step instructions about how to delete a node from a cluster using the CLI.

Attention: Attention: Before deleting a node from the cluster you should quiesce all I/O operations that are destined for this node. Failure to do so may result in failed I/O operations being reported to your host operating systems.
Prerequisites
Attention: Attention: If you are deleting a single node, and the other node in the I/O group is online, be aware that the cache on the partner node will go into write-through mode and that you are exposed to a single point of failure should the partner node fail.
Attention: Attention: If you are deleting a node, and this is the last node in the I/O group, you will lose access to all VDisks served by this I/O group. Ensure that all VDisks are not being accessed or contain data that you wish to continue to access, or ensure that they have been migrated to a different (online) I/O group.
  1. Begin by determining the VDisks that are still assigned to this I/O group:
    1. Determine the VDisks in question by requesting a filtered view of VDisks where the filter attribute is the I/O group in question. This can be done using the following command:
      svcinfo lsvdisk -filtervalue IO_group_name=<name>
      where <name> is the name of the I/O group in question.
    2. Once you have a list of VDisks, determine the hosts that they are mapped to by following the procedure called, Determining the hosts that a VDisk is mapped to.
    3. Once you have determined the hosts and are sure that you do not wish to maintain access to these VDisks proceed to 3.
    4. If you determine that some or all of the VDisks assigned to this I/O group do contain data that you wish to continue to access, you should follow the procedure called, Migrating a VDisk to a new I/O group.
  2. Before performing the SDD path removal procedure described in 3 you should power off the node that you intend to remove, unless this is the last node in the cluster. This ensures that SDD does not re-discover the paths that are manually removed before you issue the delete node request.
    Attention: Attention: If the node being removed is the configuration node, it may take a minute or so before you can perform the delete node request as the configuration node failover has to occur. If the node being removed is the last node in the cluster, the SAN Volume Controller Console may appear to hang for up to 3 minutes because you have removed the last access point to the cluster. Removing the last node in the cluster destroys the cluster. Ensure that this is what you want to do before performing this task.
    Note:Note: If you power back on the node that has been removed and it is still connected to the same fabric or zone it will attempt to rejoin the cluster. At this point the cluster will tell the node to remove itself from the cluster and the node will become a candidate for addition to this cluster or another cluster. If you are adding this node back into the cluster, ensure that you add it back to the same I/O group that it was previously a member of. Failure to do so may result in data corruption.
  3. Before deleting the node, it is essential that for each vpath presented by the VDisks you intend to remove, the SDD configuration is updated to remove the vpaths in question. Failure to do this may result in data corruption. See the IBM TotalStorage Subsystem Device Driver: User's Guide for details about how to dynamically reconfigure SDD for the given host operating system.
Steps

Perform the following steps to delete a node:

  1. Open a command prompt window.
    Notes:
    Notes:
    1. Before removing a node, be sure this is what you want to do. Any VDisks that are assigned to the I/O group that this node belongs to, will be assigned to the other node in the I/O group, that is, the preferred node will be changed. You cannot change this setting back once this has been done. Also, all VDisks will go into write-through cache mode as there is not a redundant node available to duplicate the cached information.
    2. If this is the last node in the I/O group or the last node in the cluster, you will be asked to force the delete.
    3. If this is the last node in the cluster or is currently assigned as the configuration node, all connections to the cluster will be lost. The user interface and any open CLI sessions will halt. This may result in a timeout from the command as the command cannot be completed before the node is deleted.
  2. Issue the svctask rmnode command to delete a node from the cluster. You can enter this command any time after a cluster has been created.

(C) Copyright International Business Machines Corporation 2003, 2004