General Discussion

powering down a 8.3.2 cDOT 8020 4 node cluster

Silverfuel
4,129 Views

I know this has been asked an answered several times but I am a bit new to storage and want to know what to expect. I am following the following Netapp KB on graceful shutdown: https://kb.netapp.com/app/answers/answer_view/a_id/1003836/~/what-is-the-procedure-for-graceful-shutdown-and-power-up-of-a-storage-system

 

We have two 8020 controllers each controller has 2 nodes both connected to two (CN 1610) intetrcluster switches. From the KB it sounds like for 8.3.2, all I have to do is login to the SP interface on each node and run: 'halt local -inhibit-takeover true' and wait for the LOADER prompt to appear on all nodes before shutting it down. For powerup, bring up the shelves first and then the controllers. Questions:

1) Does the above sound right?

2) Do I need to shut down the CN1610 switches?

3) During powerup, do I need to login to SP and then issue a boot command at the LOADER prompt or will it just boot into Ontapp?

 

Are there any common mistakes that people make that I can avoid by prepping in advance? 

1 ACCEPTED SOLUTION

Damien_Queen
4,072 Views
2) in case you have two node cluster:
::> cluster ha modify –configured false
This command in any case
::> storage failover modify –node * -enable false

Only then halt

View solution in original post

4 REPLIES 4

Damien_Queen
4,115 Views
1) Shutdown hosts been using your storage
2) Disable failover (and HA if needed)
3) Shutdown storage with halt command


You do not have to shutdown the switches, there is no need. But if you want, you can shut them down too after you shutdown the storage.

When you will need to power up your storage, power up switches first. Then power up storage & go to loader, through console or SP and type “boot_ontap”. After all the nodes get online, enable failover on all the nodes (and HA if needed).

Only then you can enable your hosts to interact with your storage system.

Silverfuel
4,104 Views

@Damien_Queen wrote:
2) Disable failover (and HA if needed)


Hi Damien_Queen, with regards to #2, is that a seprate command or will the 'halt local -inhibit-takeover true' in the KB take care of that? From the power-down KB, I thought that will not necessary because even the KB warning says this will not failover and if you want it then to issue a manual command.

 

 

- If that is a separate command, will this work? "storage failover modify -enabled false -node nodename"

- To disable HA, can I use: "storage failover modify -mode non_ha -node nodename"

 

Should I issue both the above commands with each nodename.

Damien_Queen
4,073 Views
2) in case you have two node cluster:
::> cluster ha modify –configured false
This command in any case
::> storage failover modify –node * -enable false

Only then halt

Silverfuel
4,052 Views

Thanks!

Public