horizonsII: Server backup causing CPU to go to 100%?

[MA70Snowman]

Title horizonsII: Server backup causing CPU to go to 100%?

Launcher Type FTB Launcher

Modpack Horizons II: Daybreaker

Modpack version 1.2.1

Have you modified the pack? Yes

Link to log file http://paste.feed-the-beast.com/view/e099d033

Details of the issue Server Specs:
CPU: Intel Devil's canyon i7 4790K @ 4 GHz
RAM: 10240 MB @ 1333 MHz
Network: 1GB
HDD: 500 GB SSD
Host: Minecraft0hosting.pro

Added to Mod pack:
Architecturecraft
Biblio Craft
Chococraft
ender compass
ender zoo
fast leave decay
grimoire of gaia
Jabba barrels
morpheus
nether core
nether portal fix
railcraft
statues
vein miner
starting inventory
bspark core
minetweaker3
asielib
MSC+

Removed:
Blood arsenal
bloodmagic
Equivilant exchange
sanguimancy
simple condensers
wiiemc

I created an offline Map and imported to the server. It was created offline to find a good seed and modify the spawn (just created a dome and set spawn)
We have 2-3 players total on the server, rarely on at same time, and no one is a huge builder (meaing all very small and not making complex builds)

We have been fine over the past 1 month, but as of three days ago... the CPU will go to 100% after our 5AM server backup, this is without fail. no one is one usually between 10pm and 4pm (most of us work) and when logging out at 10PM the CPU is usually around 35-40% which this is normal and we've always observed it that way. But after the Server backup (can see this in logs) all of a sudden we start getting server warnings every 3-7 minutes about ticks being skipped.

I've used Opis to try and track down any issues, but nothing stands out as above normal, most of us aren't even mid game and only one of us has any ty pe of automated ore processing going on. out of percautions, we've disabled a number of our automations (sugar cane farm, wheat farm, hemp farm, ethenal production via immersive engineering, tore down Aura Cascade nodes that had energy still sitting) and while we are on, the server runs fine, most of us can run at 60 fps no problem. We think the server is fixed and then the next morning find out that CPU is back at 100%, the server doesn't crash, but no one can log in, we get time out issues. Rebooting the server doesn't fix the issue, but if I restore to the 4:50 am server backup (which is only 10 minutes before the issue starts normally) I'm able to log in and opis sees nothing and CPU is back to 40%
So.. not sure what is going on...

 

[MA70Snowman]

After perusing for other solutions it appears this thread belongs under the server troubleshooting sections.. sorry about that could you please move?

 

[MA70Snowman]

here's some shots of my latest opis

 

[grandrolf]

A few questions:
Is this a dedicated server or a VPS?
How much ram is available on the machine?
How much ram is allocated for minecraft?
Is there any other software running on the machine?
Is the server launched using the .sh/.bat-files or is there any other management software?
How big is the backup?
Is there enough free disk space for the backup?
Is it a mod that does the backup or is it done with any other solution?

 

[MA70Snowman]

Is this a dedicated server or a VPS?
According to the service provider and the leve of service I paid for, I should have dedicated hardware... what that means on the backend... don't know.

How much ram is available on the machine?
10280 MB of ram

How much ram is allocated for minecraft?
I assume all of the RAM that I paid for but I don't have any scripts that I'm aware of that modify how much is allocated.

Is there any other software running on the machine?
No

Is the server launched using the .sh/.bat-files or is there any other management software?
How big is the backup?
I don't know how the server is launched, all I'm given is a GUI with limited console information.

Is there enough free disk space for the backup?
Yes my SSD is ibly using 2% of the available and I'm only keeping the 10 most recent backups

Is it a mod that does the backup or is it done with any other solution?
Another solution. Not sure what it is as I schedule the backups through the GUI

So I was troubleshooting again last night... did the same thing reloaded the most recent backup. Logged in checked opus everything looked fine. Each night I've been tearing things down and seeing what happens but I changes. Last night I removed my two capacitors from my advanced generators machine... and this morning my CPU was still fine... no idea how/why THESE capacitors would cause problems.. I have 10 HV capacitors from Immersive engineering, and there's a bunch of the mekanism energy cubes on the server. So scratching my head now

 

[MA70Snowman]

Well not sure if that was the fix.. the CPU stayed low overnight which is a first, but now almost 12-18hrs later i'm seeing the server warnings again.

 

[grandrolf]

Contact your provider, guessing something is wrong at their end if this happens about the same time as backups.

 

[MA70Snowman]

Well after wrestling with the provider, they finally agreed to move us to a new server and a new server node.. so started with a brand new server, and transfered over my previous server and.. BOOM straight up to 100% CPU.. their support is honestly useless, and me and the other players can't seem to make heads or tails of this issue. thought we had it fixed when we removed the biodiesel generator and actually had cpu drop back down to 30-40% but now the issue has returned.. I can CONSISTENTLY revert to our midnight backup.. restart the server and CPU drops to 30-40% and stays stable until the 4am server backup. I can't tell when the CPU starts to climb as I have no historical data. I'm no longer seeing the server tick warnings, but when we try and log in to the server we get time outs and when i look at the server I can see the CPU pegged at 100%. Opis doesn't show much anything different than before.

recent changes. Updated all modes to the latest 1.7.10 support versions.

 

[grandrolf]

Download your world and testrun it local and if it doesn't cap, then you might want to consider changing provider.

 

[MA70Snowman]

Download your world and testrun it local and if it doesn't cap, then you might want to consider changing provider.

I bought a test server with beastnode and tried to migrate server, sadly I couldn't get my server to transfer correctly, going to try again today. Didn't try to run locally thiight