3CX ESX VM freezes regularly

[alexschomb]

We are experiencing similar issues to this (closed) thread:
https://www.3cx.com/community/threads/3cx-vm-is-freezing-up.56780/

We setup a 3CX test VM (latest version of 3CX) on a ESX 6.5 server that successfully runs several other VMs. Unfortunately the 3CX ist regularly freezing and can only be revived by a cold restart. The VM itself is used very rarely and only two Snom phones are permanently connected to it for testing purposes. How can we run 3CX reliably?

 

[cobaltit]

@alexschomb

Is 3CX freezing or is the entire VM freezing? If the entire VM is freezing then you probably want to go to the Debian or ESX forums to see if there are any known issues or special configurations need with Debian 9 and ESX.

 

[alexschomb]

The entire VM is freezing. Can you direct me a link to the Debian / ESX subforums of 3CX? I'm having a hard time finding those. If you want to direct me to an official Debian or ESX forum then I'd like to tell you that we're running several Debian VMs on multiple ESX servers in our networks and don't experience such behavior on a such regular basis as with 3CX. Any VM might freeze at some point, but that happens once or twice a year at max! We're running the official 3CX Debian iso installation (following the official guides) for a few weeks and it froze three times already. At the current state its a matter of giving 3CX another chance or switching to a different solution. The main idea of our test installation was to gather some experience for a possible setup for one of our customers. Right now that experience is quite mixed.

 

[Lee Cramman]

I've previously had 3CX running on this exact setup as a test-bed for several months without issues. My VM should, in theory, be exactly the same as yours

Did you use the recommended settings for Debian when creating the VM? Is there anything non-standard about your virtual hardware (such as NIC choice).

Is the host and it's network card on the hardware compatibility list for ESX 6.5?

Have you tried another host in case there is a hardware fault?

Have adequate resources been allocated to the VM?

Is there anything in the logs at the time of the failure?

Have you tried a fresh install in case something in this one is corrupted?

 

[cobaltit]

@alexschomb

So you say you are running several Debian VMs which is great. Are they current? I'm not a ESX guy so I don't know the specifics but did you load the guest integration components? I know most current OS's have the integrations built-in at this point but sometimes but just a thought. Try turning off time sync with the host? Otherwise if you already run several VMs with Debian then just do a vanilla Debian install. I do that for all my installs anyways.

https://www.3cx.com/docs/manual/installing-debian-linux-pbx/

And yes I was directing you to Debian or ESX forums because at the end of the day 3CX is just an application. If you other VMs were freezing would you go to the Apache forums or whatever other application you might be running in there? Probably not. 3CX isn't using a custom kernel, isn't providing any hardware drivers, etc so if the entire VM is freezing that would makes me think it's something with the host configuration or Debian and ESX. A vanilla install would also rule that out.

Searching the forums shows some discussion about the NIC choice with the consensus being use VMXNET3 but I assume you are already doing that or using whatever you are using on your other instances.

Good luck and let us know if you figure it out.

 

[ucme4now]

Cobaltit is headed in the right direction. When using EXSI v6.5 you have change the default driver for the ethernet NIC from VMXNET3 (Default) to the Intel e1000 or you will experience 100% utilization and a lock up. I came across this a while back. It is a known problem with the linux community. You do not lose anything by making the change and you gain stability. I do not know if this has been fixed in EXSI v6.7, I have not tried it.

I'm not going to get into the technical differences between the drivers, you can research that on your own. Make the change and you will be good.

 

[Lee Cramman]

Cobaltit is headed in the right direction. When using EXSI v6.5 you have change the default driver for the ethernet NIC from VMXNET3 (Default) to the Intel e1000 or you will experience 100% utilization and a lock up. I came across this a while back. It is a known problem with the linux community. You do not lose anything by making the change and you gain stability. I do not know if this has been fixed in EXSI v6.7, I have not tried it.

I'm not going to get into the technical differences between the drivers, you can research that on your own. Make the change and you will be good.

I had wondered about that when I posted. I believe it has to do with offloading and can also be cured by disabling it but I would think the best fix is the one you suggest.