How to perform a staged (cold) reboot for Access Appliance

Article: 100063354
Last Published: 2024-02-14
Ratings: 1 0
Product(s): Appliances

Description

In some circumstances, Veritas Support may suggest performing a cold (staged) reboot of the appliance and associated storage shelves.  Since Access Appliance is always a cluster of two nodes with direct attached storage, careful consideration must be taken to avoid the possibility of data corruption.

Please follow the steps below to safely perform a staged reboot of the cluster nodes and external storage shelves:

1. Issue a shutdown command from the Access Appliance management CLISH:
> cluster shutdown all

2. Monitor the progress of the shutdown by reviewing both of the node states via the IPMI web interface and the virtual console. 

It is imperative that to not use the IPMI interface to "power off" the nodes if the operating system is still running, they must be fully powered off or have reached "system halted" state before it is safe to remove the power.  Since both nodes in the cluster have direct access to the shared storage shelves, inappropriately powering off the system too early may lead to I/O errors and data corruption.

3. Once the appliance nodes are both powered down, remove the power cables from the nodes. 

4. Remove the power cables from the first storage shelf (closest to the appliance) - this storage shelf contains the RAID controllers and should be powered off before the secondary shelves.  Please review below example of the power cabling, indicated with black lines to the PSUs with this diagram from the Veritas Access 3350 Appliance Hardware Installation Guide:

In the diagram above, there are two appliance nodes and 4 storage shelves - the first primary storage shelf contains the RAID controllers.  The RAID controller has SAS ports labeled with purple outlines, which can be used to confirm the correct component.

 

5. Remove the power cables from the remaining storage shelves (if present), in sequence, going from closest to furthest away from the appliance nodes.

6. Wait 10 minutes to allow the hard disks to fully spin down.

7. Re-apply power to the storage shelves, starting from the bottom and working up to the primary storage shelf.  Wait 5 minutes between powering on each shelf to allow the disks to spin up and initialize.

8. Once the primary storage shelf has been powered on, wait 10 minutes then reconnect power to the appliance nodes.

9. Connect to the appliances via IPMI console and power on the nodes.  Confirm that the nodes boot up successfully.  

10.  Once the virtual console shows the Access Appliance login prompt, the appliances should be able to be connected via SSH.  The management CLISH and Access Web UI may take longer to initialize but should be ready shortly after.

 

Was this content helpful?