Cisco 5520 WLC HA recovery

[cxu21]

We have 2 Cisco 5520 WLC running in HA mode. During software upgrade, for some reason, the primary finished upgrade but secondary did not, HA was broken and primary in maintenance mode and secondary became active.

Someone did quite a few clean up tasks on the secondary before realise the primary was in maintenance.

In order to bring the HA back, here is my plan and please advise if it is feasible.

1. disconnect RP ports connection.

2. on primary(WLC01) issue "config redundancy unit secondary" to change its role as secondary

3. on secondary(WLC02) issue "config redundancy unit primary" to make it as primary

4. on WLC01 issue below commands to downgrade it to the same version as WLC02

    transfer download mode tftp

    transfer download datatype code

    transfer download serverip 1.1.1.1

    transfer download path .

    transfer download filename AIR-CT5520-K9-8-10-162-0.aes

    transfer download start

5. reload WLC01 by using "reset system"

6. once reload is completed, check the version on WLC01.

7. if software version are matched, connect the RP ports.

8. if everything goes well, configuration from WLC02 will be synced to WLC01.

9. enter "config redundancy unit primary" on WLC01 to change it back as primary and enter "config redundancy unit secondary" on WLC02 to change it as secondary.

Coule you let me know if above process is the right one to follow?

Also, in step 9, will the server need reboot to make it take effect?

[Leo Laohoo]

If the RP port gets disconnected, how do you expect to fail over?  This is not possible. 

The only way is to sever all connections to the secondary unit and upgrade the firmware using the Service Port.

[cxu21]

Hi Leo,

Because the everything is working on secodary which has lower version of firmware, we prefer to downgrade the firmware on the primary server instead of upgrade the firmware on secondary.

We can leave the RP port connection untouched. 

In that case, do you think it will work?

[marce1000]

 - With such a long sequence , it is difficult to tell if everything or the path taken is completely correct  also because of the origin reason not exactly down , basically it comes down to breaking HA ; make sure both controllers are OK and healthy and build HA again , 

 M.

[Haydn Andrews]

Break HA
Treat it as an RMA, upgrade, drop config on it, join to SSO cluster
then use fail over commands to make the one u want primary.

[balaji.bandi]

check below i have documented easy step with cautious : (make sure you have backup out of the box)

https://www.balajibandi.com/?p=1808

[Rich R · Accepted solution]

Like @balaji.bandi make sure you have a config backup before you start doing any of this.
Mostly I think your plan will work - just a few comments and making it much simpler:

4.  No need to download the old software again - it should still be there on WLC01 in the backup partition - you just need to switch to it: "show boot" to confirm which version is in which partition then "config boot primary/backup" to tell it to boot off the old version on next reload.

There's no need to change the primary/secondary.  Once you have the primary running on correct version reconnect RP ports and power cycle the primary.  It should come up and sync config with secondary and then when sync is complete you can just do a "redundancy force-switchover" to switch back to running on the primary.

As @Haydn Andrews said this is essentially the same as RMA:
https://www.cisco.com/c/en/us/support/docs/interfaces-modules/wireless-services-module-2-wism2/117729-configure-wlan-00.html#anc7