Hey Folks,
This only started happening since 0716 EST on 4/18. I was seeing 50% cpu usage on my 2 Core VM immediately after startup, and the desktop was totally frozen. After a bit of troubleshooting I found that disabling the NIC allowed me to log into the desktop and use the vm, but unfortunately logs showed nothing.
This morning I rebuilt the vm, without running OS updates, to eliminate that as the cause. Without installing plex, I opened the plex.tv login page, and the system locked up exactly the same way.
I rebuilt the system again, grabbed the plex installer using wget, and installed the packege. No sooner did I get the “plex has been installed” notification, the system locked up.
I’m not sure if it’s a distro thing, but I though I’d see if anyone else is having this issue.
I apologize, it looks like it was actually 0614 EST 
I Installed Ubuntu Server 18.04.2, followed by plex, and had the exact same behavior. See SSH Below:
I see no errors in that console output.
If your VM immediately locks up, you have a hypervisor problem.
I’m beginning to agree. I spun up a system in VMware workstation, and it seems to be working just fine. For some reason this problem only affects systems running plex in my ESXi cluster.
Hosts are Lenovo M92p i5-3470’s with 32gigs of ram running ESXi 6.5
The test system where it worked was Ryzen 1600 16gigs of ram, Win10 1803, VMware Workstration 12.
It’s just strange that it ONLY locks up as soon as plex is installed, and it doesn’t matter which host it’s running on.
It only locks up for you in this way.
I have qemu and VirtualBox as my hypervisors. Neither of them ever lock up.
There is a problem there somewhere. It won’t be easy to find either. I do not envy you.
If the OS is solid and up to date, an application cannot crash the kernel unless it somehow manages to pass a null pointer to a kernel call which goes undetected. That’s a kernel bug.
1 Like
I got it working! The fix was to recreate the vm with hardware version 9. For some reason there’s some incompatibility between Ubuntu 18.04 and vmware hardware version 13.
1 Like
Thanks Chuck. Had it not been for your Kernel comment, it might have taken me a lot longer to figure this one out
1 Like
Just found this thread and was pulling my hair out. I’m running plex on an Ubuntu 18.04.2 on ESX 6.5. I finally resorted to rebuilding the VM like you, but as soon as plex was install the OS would crash. I backed off a hardware version from 13 (6.5) to 11 (6.0?). Rebuilt it again and all is well again.
1 Like
is this entirely a Vmware 13 problem?
If Plex, Need logs please.
Thanks.
I just confirmed this on my end and oodles of googling finally brought me here. I updated my ESXi host from 6.0 to 6.5 and suddenly had crashing issues on my Ubuntu 18.04 virtual server that had been fine for months. I thought it was an Ubuntu issue so I switched to Debian 9 and had the same issue, even with 100% fresh installs.
I made a new VM using 6.0 compatibility level, installed Ubuntu fresh on it, got it up to date, added the repository and ran sudo apt install plexmediaserver. It worked.
Normally the VM would 100% lock up just after the install finished (SSH refused and console would not accept input). On reboot, if you ran sudo systemctl status plexmediaserver, you’d see that it degraded, showing an exit code of 255/n/a. If you checked the logs, you’d see a bunch of SQLite3 errors saying it couldn’t find a certain table inside the database. I deleted the VM out of frustration before grabbing the logs for history, I’m sorry.
There is something in the way 6.5 works that causes issues. I didn’t try other distros or Windows but I’m willing to bet a beer you’d get similar issues with other Linux distros. I’m really hoping others that are running ESXi see this!
1 Like
It didn’t sit right to leave this without specific logs. For science, I spun up an Ubuntu 18.04.2 and a Debian 9 virtual machine on hardware version 13, got them updated and then installed plexmediaserver_1.15.5.994-4610c6e8d_amd64.deb.
Both crashed right after install and on reboot, the service was marked as failed for too many restarts. I tar’d the logs from both. journalctl just shows it trying to restart 5 times.
ubuntu18plex.tar.gz (1.0 KB)
debian9plex.tar.gz (928 Bytes)
2 Likes
Hey Everybody,
I’m glad my troubleshooting was able to help.
I just wanted to pop in and add an observation that I made that could help narrow down the issue.
I noticed that I got the exact same lock up issue, without installing plex, in the linux desktop using a vm running hardware v13 just by opening the plex login dialogue for the plex.tv webpage (in chromium and firefox). I believe that the root cause is the deployment of a new authentication process on the date mentioned on my first post. Likely, this is caused by a cpu encryption call.
1 Like
Chuck pointed me here. I was experiencing the same PMS install issues with my ESXi 6.5 and Ubuntu 18.04.2 set up. Downgraded compatibility to 6 and the issues seemed to have vanished. Thanks for all the troubleshooting guys. Took me a while to get here but hopefully others will see this thread and stop pulling their hair out like I’ve been doing over last few days.
1 Like
Just FYI, I have been running Ubuntu 18.04.2 under ESXI 6.7 (VM Version 14) for a long time with no issues crashes, so perhaps something is fixed there.
1 Like
Hey Gizmo, just popping in to tell you that I think I love you.
I’ve been fighting this for 2 weeks and finally got the right string of words together in Google to get to a helpful thread. I lowered my VM version to ESXi 5.5 and hey presto - no more crash when the install finishes!
If it’s helpful - I’d been having the same log errors as dietolead. The install would crash and my logs would just point to a missing table / corrupt database. I’ve nuked the VM’s so I can’t provide the exact logs unfortunately
1 Like
Thanks! I’m really glad I was able to help 
It’s the network driver. Switching out the virtual Nic would solve it as well.
It’s a real PITA.
Theres an incompatibility between VMware ESX 6.5 and Ubuntu 18.04 with the vmxnet3 driver.
To recap, Once again this is VMware / Ubuntu incompatbility and the VMNET3 driver?
Will they ever get it fixed? The previous workaround was to use e1000e emulation.
