Automatically restarting the LPA Grid nodes

You can run commands in the paadm tool to automatically restart the LPA grid nodes.

  1. Access the paadm tool.
  2. On the command line, run this command:
    This table shows the available <values> commands:
    Command Description
    --concurrentnodes <number of nodes>

    The number of LPA nodes to restart at a time.

    The default value is 1.

    --timetowait <time to wait>

    The number of seconds to wait for work unit threads to terminate.

    The default value is 21,600 seconds (6 hours).

    --restartwaittime <time to wait after restart>

    The number of seconds to wait before restarting the next LPA node.

    The default value is 1,200 seconds (20 minutes).

    --checktime <time to check>

    The number of seconds to wait before checking the work unit thread counts in the node

    The default value is 300 seconds (5 minutes).

    --nodelist <node list>

    A comma-separated list of LPA nodes to restart.

    The default is all LPA nodes.

    Partial host names would also work unless there are no more than one LPA node is found. For example, if the LPA node host name is HOSTNAME1 and if the --nodelist is given with HOST, it would still work but if there are 2 LPA nodes one with HOSTNAME1 and another one with HOSTNAME2 then specifying HOST to the --nodelist would result to an error.

    --skipnodelist <node list> A comma-separated list of LPA nodes to skip restarting.

    The default is none.

    This functions the same way as --nodelist for skipping the restart one or more nodes.

    --jvmlist This command would restart the nodes by jvm id. This can also be supplied with one or more jvm ids.

    If the jvmlist option is selected, then waitforspecload would automatically be set to No, which means the restart would wait for specific amount of time before restarting the next node. 20 minutes is the default and can be changed using --restartwaittime.

    --skipjvmlist This command would skip the nodes by jvmids.

    If the skipjvmlist option is selected, then waitforspecload would automatically be set to No, which means the restart would wait for specific amount of time before restarting the next node. 20 minutes is the default and can be changed using --restartwaittime.

    --yes

    Suppresses prompting to confirm the restart procedure.

    The default is to prompt to confirm the restart procedure.

    --skipwulist <dataarea>,<dataarea>:<workunit number> If some data areas or work units need to be skipped from the wait time of work units completion before moving on with restart, then this option can be used.

    Specify just the data area name if all work units in that data area can be skipped, specify the data area and work unit number with the colon symbol to skip only that specific work unit. If multiple work units or data areas need to be skipped, then create it as comma separated list.

    --changewulist <dataarea>,<dataarea>:<workunit number> If the work unit skip list needs to be changed after submitting the restart process, then this option can be used. The format is same as skipwulist and this command will replace the existing skipwulist on all the LPA nodes.
    --waitwulist <dataarea>,<dataarea>:<workunit number> This is a comma-separated list of non-production data area or work unit list to include in timetowait.

    For example: dataarea1, dataarea2:12345.

    The default value is none.

    --changewaitwulist <dataarea>,<dataarea>:<workunit number> This is a comma-separated list of non-production data area or work unit list to include in timetowait.

    For example: dataarea1, dataarea2:12345.

    The default value is none.

    --waitforspecload This option would make the restart process to wait until the current restarted LPA node to come up and load the spec repositories for all data areas before moving on to the next node to restart.

    If both restartwaittime and waitforspecload options are specified, then it will wait for either one of it to be done and move on with the next node restart.