SOLVED 1.0.1: server cpu usage

[rubigwill73]

Version:
1.0.1

What is the bug:
after 30 mins of game play/server running even if none one is on the server it will jump from 20% all the way to 90% in the server cpu and stay there until server is restarted. once restarted it gets about another 30 mins of low cpu usage then it jumps back to 90%. This happened on 1.0.0 and 1.0.1.

Mod & Version:


Paste.feed-the-beast.com log:


Can it be repeated:


Known Fix: It was aroma backup. I disabled it on my server all is well now thx!

 

[The_Damaged]

Yeah, pretty sure this is the EnderIO power cables. Cause lag bursts of about 2-3 seconds (best seen when trying to open and close chests). After extensive work today (both myself and other clients of CreeperHost) we narrowed it down to those.

This of course leaves us with a problem. If we disable them to stop people lagging up servers, what do we suggest people use as an entry-level RF transport? Takes upwards of 10 ender pearls to be able to make Extra Util energy nodes.

 

[Solax]

We have no EnderIO cables and have the same issue with 1.0.1. Server is still playable, even though capped CPU causes some block lag. Restarting helps.

 

[Duo]

Yeah, pretty sure this is the EnderIO power cables. Cause lag bursts of about 2-3 seconds (best seen when trying to open and close chests). After extensive work today (both myself and other clients of CreeperHost) we narrowed it down to those.

This of course leaves us with a problem. If we disable them to stop people lagging up servers, what do we suggest people use as an entry-level RF transport? Takes upwards of 10 ender pearls to be able to make Extra Util energy nodes.

MFR rednet energy cables are good early for energy transfer, as well. I found them easier to craft than the ender io ones, personally.

 

[It_IS_Lupus]

Version:
1.0.1

What is the bug:
after 30 mins of game play/server running even if none one is on the server it will jump from 20% all the way to 90% in the server cpu and stay there until server is restarted. once restarted it gets about another 30 mins of low cpu usage then it jumps back to 90%. This happened on 1.0.0 and 1.0.1.

Mod & Version:


Paste.feed-the-beast.com log:


Can it be repeated:


Known Fix:

Make sure your jre is up to date. Upgrading to Java 8 jre8v25 fixed it for the better part. I upgraded it at about 8 and been running since with no issues.

 

[fall1aty]

I'm pretty sure this is aroma backup i had the same problem with another modpack every time it starts to run a backup the CPU usage stuck at 90-100% and stay there.

 

[Tankh]

I'm pretty sure this is aroma backup i had the same problem with another modpack every time it starts to run a backup the CPU usage stuck at 90-100% and stay there.

Yes I also believe this is AromaBackup.

It has other issues as well. After the first backup is done (Default is every 30 minutes, which is why you see this happen after 30 minutes uptime) It never does a second update.

I checked "/Backup Schedule" and got "[§1AromaBackup§r] Next backup in 23593554 minutes."
I then tried "/Backup schedule set 60" just to increase time between backups, and the server just insta-crashed.
I will probably disable AromaBackup and instead use my own backup shellscript running outside the minecraft server. You can find the script here if you're interested: http://forum.feed-the-beast.com/threads/ftb-server-backup-script.45637/#post-798681

 

[rubigwill73]

I'm pretty sure this is aroma backup i had the same problem with another modpack every time it starts to run a backup the CPU usage stuck at 90-100% and stay there.

thx a ton

 

[fall1aty]

thx a ton

No problem

 

[GemSlate]

I had the same issue. 30 minutes after starting the server, CPU usage would jump from averaging 5-15% to 99-100% until it was restarted (new world, no EnderIO pipes or chunk loaders of any kind). I actually watched this happen as soon as AromaBackup ran.

The Fix: (other than disable)
Upgrade your Aroma1997Core and AromaBackup. Seems the versions that download with the server are not the most up to date. Aroma1997Core -1.7.10-1.0.2.9 updates to 1.0.2.12 and AromaBackup-1.7.10-0.0.0.4 updates to 0.0.0.5.

From the AromaBackup-1.7.10.-0.0.0.5 Change log
Version 0.0.0.5:
-Fixed bug with scheduled backups not starting. (This was a SEVERE bug, you may want to update).

Hope this helps others with the same issue. It seems to have fixed both of my Direwolf20 1.7.10 servers. Now if only I could get Multicraft to see the servers are up and running...

 

[Jajujo]

I tried getting rid of Aroma, removing the ender IO pipes from my base and nothing seemed to help.

I finally gave opis a shot and found the offending chunk. I started tearing out machines and rerunning opis after each to see when the update time dropped.

Turns out my major problem was from the Buildcraft Advanced Crafting Table. Removing this one block dropped the update time from 30000+ to 150! Server CPU usage also dropped from nearly 100% to around 30%.

 

[bgsteiner]

I had the same issue. 30 minutes after starting the server, CPU usage would jump from averaging 5-15% to 99-100% until it was restarted (new world, no EnderIO pipes or chunk loaders of any kind). I actually watched this happen as soon as AromaBackup ran.

The Fix: (other than disable)
Upgrade your Aroma1997Core and AromaBackup. Seems the versions that download with the server are not the most up to date. Aroma1997Core -1.7.10-1.0.2.9 updates to 1.0.2.12 and AromaBackup-1.7.10-0.0.0.4 updates to 0.0.0.5.

From the AromaBackup-1.7.10.-0.0.0.5 Change log
Version 0.0.0.5:
-Fixed bug with scheduled backups not starting. (This was a SEVERE bug, you may want to update).

Hope this helps others with the same issue. It seems to have fixed both of my Direwolf20 1.7.10 servers. Now if only I could get Multicraft to see the servers are up and running...

I was able to get my multicraft setup to see my server you just need to add this to your strat script
-Djline.terminal=jline.UnsupportedTerminal
and it recognized it perfectly it does add the method that is outputting text to the line so it can be harder to read although

 

[poxun]

I'd been getting very poor performance on my sever and after replacing Aroma backup with a linux script cron backup and while doing so realising I was running openJVM instead of Oracle Java and therefore switched. Now it runs smooth as silk.

 

[GemSlate]

I was able to get my multicraft setup to see my server you just need to add this to your strat script
-Djline.terminal=jline.UnsupportedTerminal
and it recognized it perfectly it does add the method that is outputting text to the line so it can be harder to read although

Finally got Multicraft to see my servers running. I had to make the following addition to the end of my FTBServer.jar.conf file.

[parse_log]
start=^(?P<time>?[-\d]+ )?\[?[:\d]+\]?)\s+\[?(?P<type>[^]]+)\]\:?\s+?\[[^]]+\]\s+)?(?P<line>.*)$

[parse_connect]
shortStart=logged in with entity id \d+ at

[parse_clean]
start=(^[>\r\s]+|\x1b(\[|\])[0-9;]*m?|\x08)

Multicraft now sees the server up and running as well as correctly logs the players name online.