FAS and V-Series Storage Systems Discussions
Do netapp provides any powering up procedures for FAS2020 ?
As what i can find from this article : https://now.netapp.com/Knowledgebase/solutionarea.asp?id=kb13933
It exclude the FAS2000 series, and if i used the halt -f to shutdown the controllers.
It also never mentioned what the powering up procedures would be, if i powered up Controller A after an halt -f would Controller B failed over immediately ?
See The Solution
You can follow the steps in the article you quoted.
The special thing on a 2000 Series is that head and Harddirves share a chassis, so you maybe don´t have a external shelf.
I this case you can power off the hole chassis after a gracefull shutdown with "halt". If you do have an external shelf attached,
shut the system down ("halt"), then first poweroff the external shelf and then the FAS2000.
When you want to restart the system do it in reverse order: poweron the shelf(s) wait abaout 30 Seconds for the shelf(s) to come up,
then boot the FAS2000.
Thanks for reply but my main concern is that if i'm running a cluster pair on the FAS2020, once i run half the other running controller will take over unless i use the halt -f command.
Which say i shutdown Controller A first followed by Controller B, you are saying i need to power up Controller B follow by Controller A ?
Or i can power on Controller A followed by B without any problem ?
My main concern for powering and shutdown the filer is not on the disk shelves and controller issue but more to the failover issue by the controller due to the cluster features.
Your question related to a clustered setup is actually answered within the article you are referring to:
"If the filer is clustered, enter cf disable"
This in plain English disables cluster failover, so when you power down any of the two controllers, the other one will not attempt to take over its partner identity.
And do not forget to issue cf enable after you power back both your controllers again!
So as to speak i must use the cf disable command to disable cluster first ?
and in that sense halt -f would be pretty much useless ?
Thanks and Regards,
It does the same job, yet halt -f looks far less elegant
There used to be a bug with this command (long time ago though):
View solution in original post
Noted. Thanks again !
Now i finally know what to input for my shutdown and powering up procedures... hah...