Coversion from Zone Director to vSZ Problems

[ryan_klingaman_]

I recently just converted my zone director with about 260 wireless APs over to a VSZ platfrom. I moved to the newer 3.5.1 and used the built in migration. I now have a handful of APS that are not keeping their connection with the Controller and are getting hearbeat loss.

I do have a support case open and they didn't really have an knowledge of it yet but they said I shouldn't have gone to 3.5.1 even though I find it recommended on the support website. Last thing we did was the certificate replacement At this point it seems to be stuck on about 78 remaining APS (I'm thinking that is the number of APS I'm having trouble with)

I also tried to factory default 15 APS then have them rebooted and clear the old record of out vSZ so they would register new. Sadly that doesn't seem to fix the issue.

The next things I'm going to try is theirs a newer ap patch firmware from 3.5.1.0.419 to 3.5.1.0.490. I'm hoping that the firmware flash if it does that will fix some issues. I have taken up backup though just in case this makes it worse.

What is the manual way to flash the APS kind of think that will be my next route if this doesn't work or if it gets held up on the APs that are not communicating correctly.

[michael_brado]

Hello Ryan,

    Heartbeat loss is typically a network related issue, where congestion or bottlenecks can prevent
remote APs from keeping timely synchronization with the vSZ controller(s).  Can you measure the
round-trip time (RTT) from the APs that are having trouble re-connecting?  The AP cert refresh is
important, and enabling lwapp2scg to recognize ZD APs, if not firmware converted to base Solo
code.  To try converting an AP, use 104.0.0.0.1347 from this link:
https://support.ruckuswireless.com/software/1017-zoneflex-solo-access-point-104-0-0-0-1347-ga-softwa...
When factory defaulted, you'll need DHCP option 43, or DNS to point your AP to the vSZ.  (or AP
CLI command 'set scg ip a.b.c.d' and 'reboot').

Please let us know, and there's no reason TS should have told you to stay off SZ 3.5.1 btw.

[ryan_klingaman_]

Over this weekend I recreated the virtual controller again on version 3.4.2 and this time around kept it out of the AP subnet and used DHCP option to point APS to it.

Initially everything moved over just fine but then started to act funny again though no heartbeat loss messages. I upped the resources on the VM to 8 CPUs and 24G of ram so that it went to the VM resource level with 1024 APS supported. After that everything stabilized out.

So it was either a bug in 3.5.1 or the fact that I just didn't have enough CPU and memory resources available.

Also if you don't have network issues and are like me who has had a long history of pain with heartbeat loss messages it usually has come out to CPU & memory constraints or Ruckus bugs. In the past it was on the AP resource side.