How to power down and power up one or more RecoverPoint Appliances (RPA) in the recommended sequence

How to power down and power up one or more RecoverPoint Appliances (RPA) in the recommended sequence

When this procedure should be used:

1.   Following a reboot or power up, switches, RPAs, or hosts do not see the storage LUNs.

2.   Following a reboot or power up, RPAs or hosts do not see the switches or storage LUNs.

3.   Following a reboot or power up, hosts do not see the RPAs, switches, or storage LUNs.

Power down the RPA or RPAs, one-at-a-time:

If shutting down all or multiple RPAs, it is recommended that all Consistency Groups (CG) running on those RPAs be paused or disabled, which is a disruptive operation to replication. (When shutting down only one RPA, the "preferred RPA" of the CGs running on the RPA can usually be moved to one of the other RPAs in the cluster, negating the need to pause and disable the CG.) The power off sequence of the RPAs does not matter. The power down sequence of host, switches, RPAs and arrays does matter if you are going to power off the entire environment, in which case the following order should be followed:

1.   Shutdown host/applications

2.   Disable/Pause RP CGs

3.   Shutdown RPAs (boxmgmt/shutdown) *

4.   Shutdown arrays

5.   Shutdown fabric switches

a [4] Cluster operations a [2] Detach from cluster.

Simply reverse the order to power up (at step 2 enable/un-pause the CGs).

When you power up the RPA or RPAs, do so one-at-a-time starting with RPA1 and work your way up until all RPAs are powered up.

In order to power down the RPAs, login as the boxmgmt user and choose a [5] Shutdown / Reboot operations a [2] Shutdown RPA.

In order to power up the RPAs, plug them in and/or turn them on using the power button on the front of the unit.

Unless the RPAs have been manually detached or there are issues that caused them to go into reboot regulation, then the RPAs will join the already designated cluster per the repository selection that should already be defined.

 

How to power down and power up one or more RecoverPoint Appliances (RPA) in the recommended sequence

When this procedure should be used:

6.   Following a reboot or power up, switches, RPAs, or hosts do not see the storage LUNs.

7.   Following a reboot or power up, RPAs or hosts do not see the switches or storage LUNs.

8.   Following a reboot or power up, hosts do not see the RPAs, switches, or storage LUNs.

Power down the RPA or RPAs, one-at-a-time:

If shutting down all or multiple RPAs, it is recommended that all Consistency Groups (CG) running on those RPAs be paused or disabled, which is a disruptive operation to replication. (When shutting down only one RPA, the "preferred RPA" of the CGs running on the RPA can usually be moved to one of the other RPAs in the cluster, negating the need to pause and disable the CG.) The power off sequence of the RPAs does not matter. The power down sequence of host, switches, RPAs and arrays does matter if you are going to power off the entire environment, in which case the following order should be followed:

1.   Shutdown host/applications

2.   Disable/Pause RP CGs

3.   Shutdown RPAs (boxmgmt/shutdown) *

4.   Shutdown arrays

5.   Shutdown fabric switches

* In RecoverPoint versions prior to RP 3.3, if the RPA is attached to its cluster, using the shutdown command on the RPA may sometimes result in an RPA reboot instead. In this case, 

in order to shutdown the RPA (so that it stays down), it will be necessary to first detach the RPA from its cluster. 

To do this, log in as the boxmgmt user and choose a [4] Cluster operations a [2] Detach from cluster.

Simply reverse the order to power up (at step 2 enable/un-pause the CGs).

When you power up the RPA or RPAs, do so one-at-a-time starting with RPA1 and work your way up until all RPAs are powered up.

In order to power down the RPAs, login as the boxmgmt user and choose a [5] Shutdown / Reboot operations a [2] Shutdown RPA.

In order to power up the RPAs, plug them in and/or turn them on using the power button on the front of the unit.

Unless the RPAs have been manually detached or there are issues that caused them to go into reboot regulation, then the RPAs will join the already designated cluster per the repository selection that should already be defined.