#Questions regarding Cluster Switch update (BES-53248)

1 messages · Page 1 of 1 (latest)

exotic hedge
#

Hello,

I am planning to do a Cluster Switch update on our BES-53248 Switches. No Migration, "just" an update.
However, I have come across several question I would like to ask before doing so.
Since my thoughts will be in text form, the questions itself will be inline, but starting with Q<digit> for better readability.

Starting point:
AFF A400; ONTAP 9.15.1P15; BES-53248 (no metrocluster).
BES-53248: EFOS 3.10.0.3 with RCF 1.9; Model BES-53248 (NetApp X190005).

I would like to update to "latest and greatest" EFOS (and therefore RCF).
According to this site: https://mysupport.netapp.com/site/info/broadcom-cluster-switch the latest EFOS is 3.14.0.2, with RCF 1.13.

Q1) Do I read the site correctly, that EFOS 3.14.0.2 (+ RCF 1.13) is supported for 9.15.1P12 and newer?
I find it kind of confusing, that a new EFOS can be installed within a minor update (9.15.1PX to 9.15.1PY).

Furthermore, I couldn't find any information if I can directly upgrade to EFOS 3.14.0.2 (+ RCF 1.13) or if I need to do it in steps.

Q2) Am I able to directly update to EFOS 3.14.0.2 (+ RCF 1.13) or do I need to it step-by-step starting with 3.12.0.1 (and RCF 1.10)?

I've read the instructions here https://docs.netapp.com/us-en/ontap-systems-switches/switch-bes-53248/upgrade-switch-overview.html and
updating the EFOS seems to be quite straight forward. However, I struggle with RCF.

In order not do disconnect during the update (and according to https://kb.netapp.com/on-prem/Switches/Broadcom-KBs/How_to_clear_configuration_on_a_Broadcom_interconnect_switch_while_retaining_remote_connectivity) I need to adapt the line starting with "serviceport". However, there is no wording like that in 1.13

Q3) Where do I add the line "serviceport" in RCF 1.13 to retain remote connectitivy during update?

I think that covers it for now.
Thank you!

coral spoke
#

You will see you are supported right now as is. Upgrading now is easier

#

Here’s the quick over view

  1. Make sure cluster is healthy. All lifs home,etc
  2. Set auto-revert to false on all cluster lifs. THIS IS EXTREMELY IMPORTANT if you do not want down time!
  3. Go to the first switch and shut down all ports ( this forces ONTAP to migrate all ports to the other switch while it’s updated)
  4. Follow the directions. The order sometimes matters. I usually do the code first then the rcf.
  5. At some point in the instructions it asks you to copy and paste a lot of info. This will keep network info in place so when it reboots it will come back online and you do not have to use a serial cable.
  6. Upgrade the EFOS. Follow the directions. If you do not you will end up in a boot loop that will need to be fixed by re-initing the switch!
  7. ONLY After EFOS is updated send over the rcf. There is a new process in 3.12+ such that you no longer need to modify the rcf.
  8. Apply the rcf per directions
  9. Reboot the switch
  10. Wait!!!!
  11. Make sure the switch comes back and looks good.
  12. You must wait for ONTAP to be ready. Verify all cluster ports are up and ready and not degraded. This is important.
  13. Only after the switch reboot and the ports are deemed ready, start the other switch
  14. This is always the scary part. After everything looks good you go to the second switch and shut all the ports down. ONTAP will send then go away. Even though auto revert is off it will still migrate all ports to the new switch.
    14 repeat on second switch
  15. At the end you reenable the auto revert and do a final reboot of the second switch.
    16 after some time ONTAP will send the cluster lifs back home

Follow the directions!!! Please.
I helped clean them up.

The directions are longer than they need to be for for noobs they work very well

Do not skip anything