Is there an upgrade guide that covers DR upgrade steps?

[shane.carnahan]

Looking to upgrade and running deployment of CatC from 2.3.7.6 to 2.3.7.9. It's running in DR already. Trying to find the steps to do this upgrade. Not sure if I need to break/pause DR and upgrade the witness node first and then active and backup. I've looked in the 2.3.7.9 upgrade guide and don't see it.

https://www.cisco.com/c/en/us/td/docs/cloud-systems-management/network-automation-and-management/dna-center/upgrade/b_cisco_catalyst_center_upgrade_guide/m_upgrade_to_cisco_dna_center_2_3_7_x.html

[maflesch]

Hi Shane, yes. It's in the DR admin guide.

https://www.cisco.com/c/en/us/td/docs/cloud-systems-management/network-automation-and-management/catalyst-center/2-3-7/admin_guide/b_cisco_catalyst_center_admin_guide_237/b_cisco_dna_center_admin_guide_2_3_7_chapter_0111.html

Yes, you need to pause Catalyst Center before doing any administrative task, such as upgrading.

[shane.carnahan · Accepted solution]

Thank you @maflesch I did find what I was looking for. It's under the "Administer you disaster recovery system" in the guide you linked.

Procedure

---

Step 1	Place your system on pause.
Step 2	Upgrade the appliances at your main and recovery sites to the latest Catalyst Center version (see the Cisco Catalyst Center Upgrade Guide).
Step 3	Do one of these tasks: Replace the current witness site Upgrade the current witness site
Step 4	Rejoin your system. Note   After upgrading to Catalyst Center 2.3.7 from version 2.3.4 or earlier, data migration takes place the first time a Rejoin operation is initiated. As a result, it will take longer for this operation to complete. The migration may add minutes or even hours to the completion time, depending on the amount of Catalyst Center data that's present. Keep in mind that this data migration only happens after an upgrade. This will not impact subsequent Rejoin operations. During this one-time data migration from MongoDB to fileservice, you can ignore high disk utilization critical alerts for MongoDB. These alerts are temporary. After the Rejoin operation completes, the disk consumption decreases to expected values.

[aleopoldie]

Hello @maflesch 

Is it also recommended to pause the DR before downloading the image ? Or only before activating ?

Thanks,

AL

[maflesch]

@aleopoldie you have to download the release on both sites, you cannot do that while DR is active. Because of that, you need to pause DR first.

[aleopoldie]

Makes sense. Thanks.

[JohnCaston]

We're also looking to upgrade our DR system to the same version. Did your upgrade go OK? Anything to watch out for?

[shane.carnahan]

Hello JohnCaston. Yes the upgrade went just fine following the guide and it resolved an issue I had with DR at the time. Good luck with your upgrade!

[JohnCaston]

Many thanks Shane, I appreciate you taking the time to reply.

I will go ahead and submit the change. Couple more questions, if you don't mind? How long did the upgrade take overall? Did you upgrade all nodes in parallel? Does your deployment have SD-ACCESS configured and, if so, was that impacted during the upgrade?

Kind regards,

John

[shane.carnahan]

It's been a bit since I did the upgrade but it was 2 three node clusters. I think each cluster takes a few hours to complete. The witness VM was very fast. Then re-enabling DR was pretty fast as well. No SD-Access so no impact there. Hope that helps.