Modpacks cause server to lock up

[Bitfern]

Title Modpacks lockup server

Launcher Type Server

Modpack Horizons 3 and Revelation

Modpack version 1.4.0 Horizons 3; Revelation 1.0.0

Have you modified the pack? No

Link to log file http://paste.feed-the-beast.com/view/03ca9357

Details of the issue I run the server inside under ESXI 6.5 in VM with Fedora Server 26. The Fedora install is completely clean with nothing but Java and FTB downloaded. The VM is running on 6 l5630 Threads with 8 Gigabytes of ram.

A few minutes after you join the server the VM locks up entirely and I need to hard reset it.

 

[grandrolf]

How much ram have you allocated to the VM?

How much ram have you allocated of those to the MC-server?

 

[Bitfern]

The VM has 12 GiB allocated (ignore the first post, must have looked at the wrong VM)

For Revelation I used 3072MB and for Horizons 3 I used 4096MB

 

[grandrolf]

Does the physical machine have enough free ram for the machine? (those 12 + the overhead)

How many cores have you given the VM?

Also make sure you have enough free disk space inside the VM

 

[Bitfern]

There is about 20 GB of free ram on the physical server

The VM has 6 Threads allocated

The VM has over 80 GB of free space

 

[grandrolf]

There is about 20 GB of free ram on the physical server

That should be fine then

The VM has 6 Threads allocated

If you're only using the VM for those 2 mc-servers, 6vcpu is a bit much - minecraft isn't multithreaded

The VM has over 80 GB of free space

That should also be fine

Do you get the issue if you just run 1 of the MC-servers in the VM?

Suggestion:
Run just one of them at a time - and check how much ram the MC-server actually want to use.
Java will allocate more ram for the java virtual machine (might be saying the wrong name now) - so 3+something and 4+something

 

[Bitfern]

Sorry should have clarified, I tested them separately. I only ran one of them at the time. After I found that Horizons made the VM hardlock I decided to try Revelation to see whether or not it could have been caused by something in Horizons. I have now tested Crackpack too which actually appears to be stable.

 

[grandrolf]

I've run several packs inside VM's - though using Ubuntu server on Hyper-v.

Have not tried Revelations yet, but Horizons III i've run without issues.

The issues I've had is when forgetting to do a remove cycle of backups so disk space have run out.

As the vm halts/locks it sounds like it's eating up resources too fast somewhere.

What cpu is in the host?

 

[Bitfern]

The host has 2 L5630s

I have considered switching back to proxmox or trying hyper-V as ESXI has been acting strange a few times, but I've never had a VM lock before.

 

[grandrolf]

so 16 threads in total, are you running any other vms on that host as well?
Perhaps not the fastest or best suited CPU for the task, but it shouldn't behave like this for you.

Any of the hypervisors should work, shouldn't fail like this.

Have you tried any other linux dists?

Perhaps far reached, but have you made sure there's nothing in the HW that's about to give up and therefore causing it?

 

[Bitfern]

I have four VMs running on the server:

Fedora 25 with Maria DB (Not doing anything)
Fedora 25 with Apache (Not doing anything, the apache service is not running)
Windows 10 with Adobe Media Encoder (Also sitting idle currently)

And of course the Minecraft VM with Fedora 26.

I've not tried any other linux dists, I will try setting up an Ubuntu server VM and trying there.

I have not run a burn test of the entire system in quite a while so it's possible that something is failing, but the weird part is that the lock up happens only a minute or two after joining on Horizons 3 and Revelation, but it's not happening on Crackpack. My friend has been on the Crackpack server for almost 2 hours now. Which leads me to believe that it's not caused by hardware failure.

 

[grandrolf]

Check and see if you have the /logs/fml-server-latest.log from any of the mc-servers that was running when it came to a halt - if yes, share that using paste.feed-the-beast.com - might be good to check (last one was latest.log - right?)

Which leads me to believe that it's not caused by hardware failure.

Yea, or you're just lucky right now - so hard to know

Do you have any enforced time sync going on?
(either in the vm or from host -> vm via vmware tools) - if yes, try and turn that off - see if that makes the "can't keep up"-messages occur less

 

[Bitfern]

Assuming that you by enforced time sync are refering to this. Time sync is already disabled.

Here is the log http://paste.feed-the-beast.com/view/81827686
The last one was latest.log