PeteAsking

Also just letting you guys know that 5.14 dropped support for some very old AP's like the first gen LR AP etc, so if anyone has old devices you might want to check the comparability list (and/or stay on 5.13). I will probs upgrade to 5.14 when they stop releasing a new version every 6 days because that generally means people have stopped moaning on the forums over there (and finding problems).

Not long until people auto get this update if they have it configured that way, the UI dude said: " Hello Everyone, This release has been promoted to stable. It will show up on the repositories on a later date. ( 1 or 2 weeks after promoting stable ) Regards, Glenn R." 20 hours ago now.

Guys I have been checking out 5.14 and while it seems ok, there are quite a few changes so if you have everything working well in production like me on 5.13, you should change your tag from linuxserver/unifi-controller:latest to linuxserver/unifi-controller:5.13.32-ls71 to avoid auto updating to 5.14 when it releases in a short while. Last thing I need is some stuff going offline, so might wait for a bit to see if other people have issues before jumping to the next version.

My one is currently set as "bridge" so it gets a 172 address. The honstame/IP setting is my unraids box IP.

I already posted above how to alleviate the issue of the 'memory leak'.

Sounds plausible to me.

I guess you have proved that it is possible.

It would be strange if it happened after moving the vm to the cache as a write operation to the cache is like writing to a single disk, outside of the array or anything so not sure how that could hang the system, or how it could still happen when before it was part of the array and now its happening when not part of the array if you see what I mean.

I normally just login via ssh and use the mv command to move the disk (vm must be off) then edit the xml to point to the correct location for the disk file. It doesnt take long.

If you are not using an SSD then likely VM's will be slow. Try using an SSD if you are not.

I use the command iotop. I might have had to install it with the nerdpack plugin but it shows disk activity per process when you run it.

Maybe do a step upgrade, first go to 5.10 then 5.11 then 5.12 then latest.

I have server 2019 running fine as a vm on unraid so must be something you did or in your config. No clue what you have done since you dont give much info.

Make sure you install the guest vm additions or vms will be slowish. Ie, disks shouldn’t be set to virtio and not ide or something if you have done it correctly.

This happened to me once when I edited the xml incorrectly. I rebuilt the vm very carefully and no longer had to change the res 2 times each boot.

Dont know but if you backup the entire unifi folder in your appdata you can always revert even if you break it.

Not really a resolution, but opnsense is pretty similar and seems to have better driver support so if you cant fix it you could switch to that. Thats what I ended up having to do when I eventually gave up on pfsense on unraid.

Have you tried changing the network card to a e1000-82545em in the xml and retesting on the vm?

You browse to http://172.17.0.6:8222/ in a new browser window. There is no option in the unraid interface.

Ok well regarding the 'memory leak' I have provided 2 potential ways to mitigate this (1 - set limit on docker, and 2 - modify the system.properties in order to prevent the database growing without running any cleanup operations). Between these 2 fixes there shouldnt be any further issues. Im on the latest version and works fine. A 2GB limit seems sufficient for the docker image, although 4GB is what Unifi themselves allocate their controllers so if you have the RAM to spare, I would use 4GB if possible. If you feel 2, or 4 GB is too much memory for a docker, well thats just how unifi coded it.

@Hoopster did you try?

We are living in the future!

I cannot imagine anything good happening to you if you use DMZ mode on your router and add your unraid server into it. It sounds like a disaster.

Actually I thought of a 3rd way : upnp on your router could be enabled possibly.

There are only 2 possible ways this can happen (assuming you are not on a vpn etc on your phone as you claim). 1) You have assigned your unraid box a public IP on a range you own. 2) Your router forwards ports to your unraid box. Pick one and look there.