ONTAP Discussions

cf takeover during a 7.3.2 to 8.0 7-mode non-disruptive upgrade (NDU)

michaelmattes
2,909 Views

How do you treat clustered filers when upgrading from DOT 7.3.2 to 8.0 7-mode?

When we upgraded from DOT 7.2.X to 7.3.2 we followed this approach to remain non-disruptive (We have two filers (named A and B) set up in a failover cluster):

  1. software install and download on filer A and B
  2. cf takeover on filer A
  3. reboot filer B
  4. cf giveback on filer A
  5. cf takeover on filer B
  6. reboot filer A
  7. cf giveback on filer B

Has anyone done this approach (or something similar) for their 7.3.X ot 8.0 7-mode upgrade?  I did see that a NDU is possible in the DOT 8 upgrade guide, just not how to perform it.

1 ACCEPTED SOLUTION

scottgelb
2,909 Views

The upgrade guide at http://now.netapp.com/NOW/knowledge/docs/ontap/rel801/pdfs/ontap/upgrade.pdf gives the deteailed instructions and all the documentation for 8.0.1 is at http://now.netapp.com/NOW/knowledge/docs/ontap/rel801/   I would first run upgrade advisor from your autosupport page and also check to make sure you meet the requirements.  The basic process is what you listed except you need to run "cf takeover -n" since the ONTAP kernel is a major update then you have to run halt on the partner node to make the takeover complete.  The first takeover/giveback doesn't need this but the second time.  The procedure is detailed in the upgrade guide starting on page 73  "Upgrading HA configurations from an earlier release family nondisruptively"

View solution in original post

1 REPLY 1

scottgelb
2,910 Views

The upgrade guide at http://now.netapp.com/NOW/knowledge/docs/ontap/rel801/pdfs/ontap/upgrade.pdf gives the deteailed instructions and all the documentation for 8.0.1 is at http://now.netapp.com/NOW/knowledge/docs/ontap/rel801/   I would first run upgrade advisor from your autosupport page and also check to make sure you meet the requirements.  The basic process is what you listed except you need to run "cf takeover -n" since the ONTAP kernel is a major update then you have to run halt on the partner node to make the takeover complete.  The first takeover/giveback doesn't need this but the second time.  The procedure is detailed in the upgrade guide starting on page 73  "Upgrading HA configurations from an earlier release family nondisruptively"

Public