ESXi Question

[LIGISTX]

OK, FreeNAS friends! I have determined what causes the issue, now hopefully there is an easy solution.

I reset my FreeNAS guest to factory, restarted it, and was able to mount my old ZFS filesystem totally file. It showed the directory structure no problem at all in the WebUI! Then I once again uploaded my config file and the above screenshots once again resulted. "Unable to load /api ect ect". Is this an expected result now that I am under a virtualized environment? Do configs not like that? Is there a line of code that can be edited to fix it?

 

[LIGISTX]

Ok, well, I guess if anyone comes along, all is up and running!!!

I figured I could mount the ZFS volume and recreate all users/settings/ect faster than I could get an answer especially considering I can load the broken VM and the real VM side by side and just go screen by screen. FreeNAS is up and running fine (seemingly, I mean, should be fine I think and seriously hope). I have a few ubuntu server VM's spun up, one with Plex and Tautulli, one with Syncthing, and a few others created but not started yet. Right now I am already murdering my RAM and CPU with Plex and Syncthing's initial sync's just hammering it. I am starting to understand why 64 GB of RAM is in so many peoples boxes/homekits....

I currently gave FreeNAS 16 GB of RAM, although I may even drop that back to 12 to see if I can squeeze a little more for other VM's. Right now I am sitting at 40% arc hit rate, but once again not many people hit this thing, and not much happens at once.

I still plan to have a VM for openVPN, nginx, and DDNS. Theoretically that VM won't need much RAM, but we will see.

Thanks for all the help!!! I am stoked to have this up and running, not being tied to freebsd jails is refreshing (although I know next to nothing about debian), its still nice to have a much larger source of knowledge and resources to pull from.

 

[LIGISTX]

Ok, just an update. I gave freenas 16 GB of RAM and that is with syncthing jail still running on it since having directed mounted storage seemed like a better idea for my use case...

Anyways, I hit one of my ubuntu servers with basically the max possible load it could have. It was downloading multiple concurrent streams via lftp from a source, and I had my phone and computer watching plex streams. The Ubuntu download was writing to a freenas directory mounted via SMB in Ubuntu, so everything was being exercised. CPU was hit from the virtual network in multiple ways with a few hundred Mbps on it, freenas was writing and serving media to and from multiple places with multiple files, and everything was totally fine.

I am showing just shy of 50% ARC hit rate in freenas, and nothing slowed down at all and I let the "test" run for a good few minutes. I ran htop in Ubuntu and I assigned it 3 GB in ESXi and it showed only using ~500MB. One question I do have, I don't have all the RAM allocated to Ubuntu like I had to do for FreeNAS to get passthrough to work, but ESXi still shows Ubuntu as using 3 GB even though htop only shows sub 1GB. Does that just mean the OS has a potential to use up to 3GB so ESXi is just telling me that number? Either way, seems totally fine and I still have a few GB left over. I do have another VM to roll out, but I currently show 7 GB free in ESXi webUI, thus more than enough!

And the CPU seems fine enough. I guess the ULTIMATE test would to also hit syncthing with a big file sync operation, while Ubuntu was lftping away and Plex was running, but honestly the likely hood of that is low. That said, I did see about 90% CPU utilization while this was going, and both FreeNAS and Ubuntu have 2 virtual CPU's, so I think theoretically it could have maxed out the CPU if it needed to in my test.

So far, it all seems spectacular. I got another issue with syncthing figured out today as well, i needed a umask for it to work as I wanted, and that worked!

As of right now, things seem GREAT! Don't even see a need for a better CPU at this point in time.