Forum


Notice to all users

We are migrating towards a new forum system located at community.teamspeak.com, as such this forum will become read-only on January 29, 2020

Results 1 to 12 of 12
  1. #1
    Join Date
    October 2014
    Posts
    12

    High load on my linux server

    Hi guys, i'm having problems with my teamspeak3 server, the load is useully around 0,00 - 0,1 but very often it peaks to ~0,8 - 1 which sometimes makes the server lag. Im running several small teamspeak3 server on my server but there is a very low amount of users connected ~50.

    The cpu is not the issue only the load here are some screenshots

    Click image for larger version. 

Name:	8Ef6h3c.png 
Views:	117 
Size:	147.7 KB 
ID:	13657

    Click image for larger version. 

Name:	d156e93d8b6f42515a16c4faf32e0879.png 
Views:	83 
Size:	25.8 KB 
ID:	13658

    Regads,

  2. #2
    Join Date
    February 2012
    Location
    Germany
    Posts
    577
    The load is in percent, which can go from 0 to 100. A load of 1% from the Teamspeak server is perfectly fine and no cause for lags.

  3. #3
    Join Date
    October 2014
    Posts
    12
    Quote Originally Posted by Schlumpi View Post
    The load is in percent, which can go from 0 to 100. A load of 1% from the Teamspeak server is perfectly fine and no cause for lags.
    Im pretty sure that for example 0,8 load is 80%...

    http://blog.scoutapp.com/articles/20...-load-averages

    Only lags at the load peaks tho

  4. #4
    Join Date
    February 2012
    Location
    Germany
    Posts
    577
    The column is titled %CPU, so the value is %, so 1 means 1%. If you need a comparison, execute "gzip < /dev/zero > /dev/null" and look on the CPU value in another session.

  5. #5
    Join Date
    October 2014
    Posts
    12
    Quote Originally Posted by Schlumpi View Post
    The column is titled %CPU, so the value is %, so 1 means 1%. If you need a comparison, execute "gzip < /dev/zero > /dev/null" and look on the CPU value in another session.
    Im not talking about the %CPU

    Click image for larger version. 

Name:	zNLCdql.png 
Views:	119 
Size:	148.1 KB 
ID:	13659

  6. #6
    Join Date
    February 2012
    Location
    Germany
    Posts
    577
    Well, the load average shows the number of processes that are in the running state (and not sleeping) within a given time frame. Since Linux is a multitasking operating system, even more than 1 processes can be running fine in parallel without any problem. They disturb themselves only, if they steal RAM, CPU or I/O from each other. Your RAM is fine. Since your CPU load is about 1% (the CPU is idle 99% of the time), you don't have any CPU load problem as well. I/O load is not shown here. Disk I/O is probably not the issue, since disk I/O from Teamspeak is very low. This leaves us with network I/O, and this is something Teamspeak is heavily depending on. I suggest you investigate how your network performs.

    There are multiple tools to measure I/O traffic. iostat is for disk I/O, in package sysstat. For network I/O, there is iftop.
    iostat can also measure disk latency, but iftop cannot measure network latency - iftop only measures throughput. If you want to measure network latency, run remote tests, like long running ping to your system.

  7. #7
    Join Date
    October 2014
    Posts
    12
    Quote Originally Posted by Schlumpi View Post
    Well, the load average shows the number of processes that are in the running state (and not sleeping) within a given time frame. Since Linux is a multitasking operating system, even more than 1 processes can be running fine in parallel without any problem. They disturb themselves only, if they steal RAM, CPU or I/O from each other. Your RAM is fine. Since your CPU load is about 1% (the CPU is idle 99% of the time), you don't have any CPU load problem as well. I/O load is not shown here. Disk I/O is probably not the issue, since disk I/O from Teamspeak is very low. This leaves us with network I/O, and this is something Teamspeak is heavily depending on. I suggest you investigate how your network performs.

    There are multiple tools to measure I/O traffic. iostat is for disk I/O, in package sysstat. For network I/O, there is iftop.
    iostat can also measure disk latency, but iftop cannot measure network latency - iftop only measures throughput. If you want to measure network latency, run remote tests, like long running ping to your system.
    http://serverbear.com/benchmark/2016...iEJ22Zuoqe0EJY Im with RamNode and they have from my experience the best network in europe There must be a way to see what process or so is causing the high load right?

  8. #8
    Join Date
    November 2006
    Location
    Kent - UK
    Posts
    67
    Click image for larger version. 

Name:	31d70b2f10.png 
Views:	90 
Size:	75.3 KB 
ID:	13666
    your averages are quiet high imo.. guess it depends what people are using it for? do you allow file transfers? check to see whats being uploaded/downloaded


    most likely your webserver causing the load differences... thought it was just a server for TS,

  9. #9
    Join Date
    October 2014
    Posts
    12
    Quote Originally Posted by Martyn View Post
    Click image for larger version. 

Name:	31d70b2f10.png 
Views:	90 
Size:	75.3 KB 
ID:	13666
    your averages are quiet high imo.. guess it depends what people are using it for? do you allow file transfers? check to see whats being uploaded/downloaded


    most likely your webserver causing the load differences... thought it was just a server for TS,
    its just apache2 installed not running anything else on it.. So odd that there only are some spikes a couple of times a day :c Havent found what process is causing it yet though. Killed the apache-2 process will see if that changes anything.. No uploads/downloads either btw

  10. #10
    Join Date
    December 2004
    Location
    RF
    Posts
    3,008
    1. http://lmgtfy.com/?q=load+average+and+how+to+read+it
    2. Sort by "TIME+", that's a better indication of system load caused by the running daemons. Especially those like TS3, which doesn't fork.

  11. #11
    Join Date
    October 2014
    Posts
    12
    Quote Originally Posted by ANR Daemon View Post
    1. http://lmgtfy.com/?q=load+average+and+how+to+read+it
    2. Sort by "TIME+", that's a better indication of system load caused by the running daemons. Especially those like TS3, which doesn't fork.
    Second searchresult is literally what i posted a couple of posts ago

  12. #12
    Join Date
    December 2004
    Location
    RF
    Posts
    3,008

    ToDo List

    Did you _READ_ it before posting? Did you understand what you read?

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Teamspreak 3 Server High CPU load
    By pafreakflorian in forum Linux / FreeBSD
    Replies: 9
    Last Post: May 17th, 2013, 10:13 AM
  2. Replies: 5
    Last Post: November 15th, 2011, 04:24 AM
  3. [Server] Very High CPU-Load
    By InComplete in forum Linux / FreeBSD
    Replies: 4
    Last Post: June 13th, 2011, 06:28 PM
  4. TS Server creating high cpu load
    By Sprenger120 in forum Bug Reports [EN/DE]
    Replies: 9
    Last Post: October 14th, 2010, 07:47 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •