Update Process - Detailed Overview?

[Paul Hutchings]

Does anyone know if there is a PDF or document somewhere with a detailed overview of the P4000 update process?

I need to update our management group, which consists of six nodes + FOM.

When I've done it before all I know is that it goes off and does "stuff" for about an hour, but I would dearly like to know more about exactly what steps/precautions the update process takes to ensure that a node has updated and rebooted successfully before moving on to the next one?

Can you ever have a situation where a cluster is unavailable because both nodes are being updated/rebooted at once?

[bryan_hadzik]

With version 9+ they hide all the peticulars of the ugprade. The inteligence is there to keep the cluster healthy though. If it fails on one node, it stops, rather than continuing on, and rebooting another node causing data availability issues.

What it is doing under the covers is this:

First determine the order. Start with nonmangers, then managers that aren't the VIP, VIP last.

For each module, upload the upgrade file, let it upgrade/reboot if needed. Let all volumes resync before starting on the next one.

The new method even does patches after the fact, and everything in the right order.

[Bart_Heungens]

Hi,

The explanation of Bryan is correct. The intelligence is there to do everything step by step to keep all volumes online... If this is not possible, U will get an alert...

On top of that I can add that inside the CMC under Help - Preferences - Upgrades you can set the Upgrade Selection Method to Advanced (default Normal) so that you have more possibilities in the upgrade process...

I sugest to keep the default but it is good to know that the option exist...

Kr,

Bart