Upgrade Trouble 11.26.3.6 to 12.8.2.2

[karsten_berghau]

Hi,

i followed the documentation to upgrade our On Premise Hivemanager 11.26.3.5 to 11.26.3.6, exported the VHM, shut down the VM, rolled out the 12.8.2.2 OVF and tried to import the VHM.

I got some trouble in this.

The import allways shows errors because the APs SN are duplicated because they auto connect to the new VM but after severeal tries of deleteing the APs and then importing it seems most settings where imoprted. The Policy and the objects are there. I had to readd additional accounts (guestmanager, 2nd admin). The maps are imported but devices are not assigned.

I did not find information if the OVF settings have to be altered in vCenter vApp section like in the old version for VMWare Tools, IP settings, FQDN...?

The test connection to License Server shows error because of missing proxy configuration but we do not have a http proxy.How do i get this to work?

The entitlement keys can not be reused. Seems i have to get new ones?

SMTP Server now needs an authenticated account? We used an internal recieve connector with anonymous authentication before but without username / password it does not work anymore?

Maybe someone can give me some help. The documentation did not provide all the information needed to do a clean upgrade.

TIA

Karsten

PS: sorry for typos. My keyboard is sometimes bouncing on some keys 😞

[karsten_berghau]

So, to bring this thread to an end, here is a short update.

Needed 5 tries of deploying a new VM with 12.8.2.2 and importing the config from 11.26.36. The only way of import without errors was to disconnecct all APs from the network during the import.

I tried to give the new VM all network config with the OVF vApp options but i did not work.

So i had to set it after the first boot under the ...:3000 admin interface and then reboot to have a working certificate and get rid of the DHCP IP. Documentation should be clearer at this point.

Still wait for my keys to be reinitialized. Opened a ticket at my reseller. Running in trial mode right now.

The new Version seems to be much slower, unresponsive, dashboard not showing connected clients,... and consuming tons of resources, and this for managing 8 APs? If this is the way of aerohive on premise VAs, we will look for another product. Cloud is no option for us. My VA has 24 GB RAM and 8 Cores and feels like an old i386 trying to run Windows 10.

With the new version we find way more problems of devices not reconnecting after they where in standby. Waking up an iPhone shows it disconnected from the WIFI and sometimes takes 2 or 3 manual connect tries to bring it back to the network.

We are only using PSK and PPSK. No Radius, AD, Cloud ID Manager, etc.

At the moment we are not very happy with the new Version 12.x.

cu

[samantha_lynn]

I would recommend opening a support ticket to dive in to this subject on that level, but in general we won't be able to alter what OS devices need what login procedure past the SSID authentication level. I also do recommend redeploying and importing the VHM data again.

[karsten_berghau]

Hi and thanks for the answer.

I had no problems with importing devices (my APs). The import of the VHM Export file from the old VA stopped because it found duplicate SN. That was caused by the APs already connecting to the new VA while i was starting the import.

Right now the new VA is running with a trial license. All APs are connected and managed, new HiveOS has been pushed to the APs.

The Network Policy and all Objects seem to be imported.

Clients can connect, but all iOS devices need to reenter the password for the same SID they used before. Windows Laptops dont need to reenter the password. I have no idea why the iOS devices can not connect again. This coulld result in some extra work because we have over 100 iPads that are managed by IT for one special appllication and where we have to reenter the password to join them back to WIFI.

My iPhone had no trouble connecting, but it gets pushed the WIFI config from our MDM (BlackBerry UEM) over the provider network, so it gets the password on this way.

Whats about the Proxy Server settings for license server? Do i need it when i´m not using any cloud hosted aerohive services? Why does it not connect without a proxy and port configured?

When everything seems to work, would you although suggest to roll out the new VM again and retry the import of the VHM data? I´m not 100% sure everything is the way it should.

Whats about the vAPP OVF settings? Are they still needed?

I will get in contact for the keys to get reinitialized when I´m sure the actual 12.8.2.2 VM stays in production environment and the old shutdown VM does not need to be powered on again.

Btw: it was not the best idea to change the SMTP to require auth. There are a lot of business environments where you do not want the HM to send auth but do it with seperated recieve connectors without creating extra user accounts on your exchange cluster just to let a VM authenticate to send some emails. 

And a little bit more info in the documentation would be helpfull and timesaving when settings have changed.

cu

Karsten

[samantha_lynn]

For the devices not importing correctly, we've had luck in the past with changing the IP address on the new deployment to something different temporarily, importing the devices, and then changing the IP back to what it was.

In general, you'll want to open a technical support case for help with most of this. I can tell you that you don't need new keys, if the old instance will no longer be running at all we can simply reinitialize your old ones for you via technical support. Also, we did change our SMTP settings to require authentication now.