PMALLOC_DOUBLE_FREE crash after upgrading to 8.5.120.0

[bvj197222]

I have two 5508 Controllers in an SSO-cluster. I upgraded from 8.3.140 to 8.5.120.0, and one of the Controllers crash With the Message PMALLOW_DOUBLE_FREE - see encl. I was able to revert to 8.3.140, rebooting both Controllers back to previous Version. However there's a bug in 8.3.140 so we need the upgraded version. With the 8.3.xx all of Our switches are being flooded With false duplex mixmatch-errrors on switchports Connected to 28xx AP's;

Mar 21 08:30:25.293: %CDP-4-DUPLEX_MISMATCH: duplex mismatch discovered on GigabitEthernet3/0/6 (not half duplex), with AP1 GigabitEthernet0 (half duplex).
Mar 21 08:30:25.286: %CDP-4-DUPLEX_MISMATCH: duplex mismatch discovered on GigabitEthernet3/0/6 (not half duplex), with AP1 GigabitEthernet0 (half duplex). (NUSBsw01-3)
Mar 21 08:30:33.689: %CDP-4-DUPLEX_MISMATCH: duplex mismatch discovered on GigabitEthernet1/0/23 (not half duplex), with AP5 GigabitEthernet0 (half duplex).
Mar 21 08:30:42.270: %CDP-4-DUPLEX_MISMATCH: duplex mismatch discovered on GigabitEthernet2/0/16 (not half duplex), with AP6 GigabitEthernet0 (half duplex).

The 8.5.120 fixes this error. Any help appreciated. If I have to reinstall the crashed Controller, any links to how I do this the best way on a Controller part of SSO? The crashed Controller (all though now working on 8.3.140) is not primary in SSO.

[Leo Laohoo]

What is the controller doing now? Is it on a boot-crash-loop?

[bvj197222 · Accepted solution]

It was in a boot crash loop. I entered setup using console cable and selected 8.3.140 as default image instead and now the Controllers are up and running on the previous Version. And the Duplex error Messages is flooding the switch-logs again..

[bvj197222]

I did an Upgrade to 8.5.140, and the SSO-cluster came back online perfectly. 8.5.120 has been deferred now as well, so that explains it..