OpenVZ Forum


Home » General » Support » High Load Issues
High Load Issues [message #12615] Fri, 04 May 2007 11:56 Go to next message
aeterna is currently offline  aeterna
Messages: 3
Registered: May 2007
Junior Member
I have been experiencing high load on my HN. Sometimes the load get's as high as 100+. I've narrowed it down to inproper configuration files for our VE's however I was hoping someone could just confirm my suspicions. I just started reading up on everything the wiki has but its still alot of information that im catching up on..

Here is an example of what one of our VE's UBC parameters look like:

# UBC parameters (in form of barrier:limit)
# Primary parameters
AVNUMPROC="80"
NUMPROC="385"
NUMTCPSOCK="160"
NUMOTHERSOCK="2147483647"
VMGUARPAGES="12288:2147483647"
# Secondary parameters
KMEMSIZE="16777216:16777216"
TCPSNDBUF="4880000:8000000"
TCPRCVBUF="4880000:8000000"
OTHERSOCKBUF="132096:336896"
DGRAMRCVBUF="1058384:1058384"
OOMGUARPAGES="16144:2147483647"
# Auxiliary parameters
LOCKEDPAGES="32:32"
SHMPAGES="8192"
PRIVVMPAGES="351072:422978"
NUMFILE="14280"
NUMFLOCK="100:110"
NUMPTY="36"
NUMSIGINFO="256"
DCACHESIZE="5916672:6163200"
PHYSPAGES="0:2147483647"
NUMIPTENT="2147483647"


So if i take 12288*4096 = 50331648 bytes = 48 MB = This is my guaranteed memory limit?

while my burstable is something like 1.3GB, correct?

[Updated on: Fri, 04 May 2007 11:57]

Report message to a moderator

Re: High Load Issues [message #12616 is a reply to message #12615] Fri, 04 May 2007 12:49 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member

by "load" I guess you mean "load average" reported by top and uptime commands, right?

Load average can be due to a number of reasons:
- you have set cpulimits to VE
- too much disk I/O (many tasks in D state waiting for I/O)
- some bug/deadlock (some task is always in D state)
- many CPU-bound applications running (tasks in R state)

Load average itself is a sum of number of tasks in R and D states. So you need to investigate task states via:
# ps axf
# ps axf -o pid,wchan,comm
# Alt-SysRq-T


http://static.openvz.org/userbars/openvz-developer.png
Re: High Load Issues [message #12618 is a reply to message #12616] Fri, 04 May 2007 13:06 Go to previous messageGo to next message
aeterna is currently offline  aeterna
Messages: 3
Registered: May 2007
Junior Member
yea, im referring to the what im seeing in my top commands. But i think whoever initially set this up didn't set the basic configuration for our vps's correctly. Initially we had cpulimits set to a certain value however were instructed to set them all to 0 or to unlimited, which didnt make any sense to me.

Was I correct with my equation about the VMGUARPAGES/PRIVVPAGES variable?

btw thanks for the reply and help
Re: High Load Issues [message #12625 is a reply to message #12615] Fri, 04 May 2007 13:43 Go to previous messageGo to next message
rickb is currently offline  rickb
Messages: 368
Registered: October 2006
Senior Member
High load on the HN means many procs in the runqueue. It does not necessarily mean 0% idle cpu (as it usually does in a non HN/vps server). many procs and idle cpu occurs when VEs have a low cpu limit and they are sustaining that limit. So, does your HN have idle cpu while having 100 load?



-------------
Common Terms I post with: http://wiki.openvz.org/Category:Definitions

UBC. Learn it, love it, live it: http://wiki.openvz.org/Proc/user_beancounters
Re: High Load Issues [message #12628 is a reply to message #12618] Fri, 04 May 2007 15:19 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member

not sure what you mean by "which didnt make any sense to me".
do you have cpulimits or not? Smile plz do the following:
# cat /proc/fairsched2
and post it here...

yes, your math is correct.
you can do:
# vzctl set VEID --vmguarpages 48m --save
instead of calculating pages.


http://static.openvz.org/userbars/openvz-developer.png
Re: High Load Issues [message #12633 is a reply to message #12615] Fri, 04 May 2007 19:44 Go to previous messageGo to next message
aeterna is currently offline  aeterna
Messages: 3
Registered: May 2007
Junior Member
cat /proc/fairsched
Version: 2.6 debug
veid id parent weight rate tasks run cpus flg ready start_tag value delay
0 0 0 14 0 225 226 1 .. 1 94611247807425 0 0
0 2147483647 0 14 0 225 226 1 .. 1 94611247807425 0 0
0 90 0 31 4096 40 0 0 L. 0 94613261814411 17364433920 4373176203264498
90 0 90 31 4096 40 0 0 L. 0 94613261814411 17364433920 4373176203264498
90 1 0 31 4096 40 0 0 L. 0 94613261814411 17364433920 4373176203264498
0 1 0 500 0 33 0 0 .. 0 94613389381069 0 0
1 0 1 500 0 33 0 0 .. 0 94613389381069 0 0
1 1 0 500 0 33 0 0 .. 0 94613389381069 0 0
0 31 0 31 4096 57 0 0 L. 0 94589523243940 17364433920 4372836721617697
31 0 31 31 4096 57 0 0 L. 0 94589523243940 17364433920 4372836721617697
31 1 0 31 4096 57 0 0 L. 0 94589523243940 17364433920 4372836721617697
0 83 0 31 4096 45 0 0 L. 0 94616983438502 17364433920 4372997905031502
83 0 83 31 4096 45 0 0 L. 0 94616983438502 17364433920 4372997905031502
83 1 0 31 4096 45 0 0 L. 0 94616983438502 17364433920 4372997905031502
0 88 0 31 4096 60 0 0 L. 0 94583732263958 17364433920 4373090581760476
88 0 88 31 4096 60 0 0 L. 0 94583732263958 17364433920 4373090581760476
88 1 0 31 4096 60 0 0 L. 0 94583732263958 17364433920 4373090581760476
0 91 0 15 4096 42 0 0 L. 0 94580675582920 17364433920 4373154798021511
91 0 91 15 4096 42 0 0 L. 0 94580675582920 17364433920 4373154798021511
91 1 0 15 4096 42 0 0 L. 0 94580675582920 17364433920 4373154798021511
0 94 0 31 4096 51 0 0 L. 0 94612780609535 17364433920 4373057796476552
94 0 94 31 4096 51 0 0 L. 0 94612780609535 17364433920 4373057796476552
94 1 0 31 4096 51 0 0 L. 0 94612780609535 17364433920 4373057796476552
0 99 0 31 4096 42 0 0 L. 0 94613468898127 17364433920 4373211672662200
99 0 99 31 4096 42 0 0 L. 0 94613468898127 17364433920 4373211672662200
99 1 0 31 4096 42 0 0 L. 0 94613468898127 17364433920 4373211672662200
0 100 0 31 4096 46 0 0 L. 0 94610065335620 17364433920 4372522768509284
100 0 100 31 4096 46 0 0 L. 0 94610065335620 17364433920 4372522768509284
100 1 0 31 4096 46 0 0 L. 0 94610065335620 17364433920 4372522768509284
0 103 0 31 4096 42 0 0 L. 0 94610619084712 17364433920 4373317617056216
103 0 103 31 4096 42 0 0 L. 0 94610619084712 17364433920 4373317617056216
103 1 0 31 4096 42 0 0 L. 0 94610619084712 17364433920 4373317617056216
0 104 0 15 4096 47 0 0 L. 0 94591515911221 17364433920 6005840073060709
104 0 104 15 4096 47 0 0 L. 0 94591515911221 17364433920 6005840073060709
104 1 0 15 4096 47 0 0 L. 0 94591515911221 17364433920 6005840073060709
0 169 0 15 4096 70 0 0 L. 0 94618630219252 17364433920 4373523471763362
169 0 169 15 4096 70 0 0 L. 0 94618630219252 17364433920 4373523471763362
169 1 0 15 4096 70 0 0 L. 0 94618630219252 17364433920 4373523471763362
0 179 0 15 4096 42 0 0 L. 0 94600751818083 17364433920 4373576791826538
179 0 179 15 4096 42 0 0 L. 0 94600751818083 17364433920 4373576791826538
179 1 0 15 4096 42 0 0 L. 0 94600751818083 17364433920 4373576791826538
0 479 0 31 4096 69 0 0 L. 0 94581043769286 17364433920 4372214340662038
479 0 479 31 4096 69 0 0 L. 0 94581043769286 17364433920 4372214340662038
479 1 0 31 4096 69 0 0 L. 0 94581043769286 17364433920 4372214340662038
0 199 0 31 4096 42 0 0 L. 0 94579551579246 17364433920 4373682218820909
199 0 199 31 4096 42 0 0 L. 0 94579551579246 17364433920 4373682218820909
199 1 0 31 4096 42 0 0 L. 0 94579551579246 17364433920 4373682218820909
0 480 0 31 4096 68 0 0 L. 0 94606770318073 17364433920 4925320687894040
480 0 480 31 4096 68 0 0 L. 0 94606770318073 17364433920 4925320687894040
480 1 0 31 4096 68 0 0 L. 0 94606770318073 17364433920 4925320687894040
0 211 0 15 4096 45 0 0 L. 0 94613455871164 17364433920 4373716229002812
211 0 211 15 4096 45 0 0 L. 0 94613455871164 17364433920 4373716229002812
211 1 0 15 4096 45 0 0 L. 0 94613455871164 17364433920 4373716229002812
0 92 0 31 4096 48 0 0 L. 0 94617287409219 17364433920 4373258475828748
92 0 92 31 4096 48 0 0 L. 0 94617287409219 17364433920 4373258475828748
92 1 0 31 4096 48 0 0 L. 0 94617287409219 17364433920 4373258475828748
0 202 0 31 4096 48 0 0 L. 0 94610614577033 17364433920 8191263418904807
202 0 202 31 4096 48 0 0 L. 0 94610614577033 17364433920 8191263418904807
202 1 0 31 4096 48 0 0 L. 0 94610614577033 17364433920 8191263418904807
0 89 0 15 4096 49 0 0 L. 0 94613011455904 17364433920 4377486407989553
89 0 89 15 4096 49 0 0 L. 0 94613011455904 17364433920 4377486407989553
89 1 0 15 4096 49 0 0 L. 0 94613011455904 17364433920 4377486407989553
0 93 0 31 4096 49 0 0 L. 0 94619171211298 17364433920 0
93 0 93 31 4096 49 0 0 L. 0 94619171211298 17364433920 0
93 1 0 31 4096 49 0 0 L. 0 94619171211298 17364433920 0
0 87 0 7 4096 45 0 0 L. 0 94607604219838 17364433920 7564054593336869
87 0 87 7 4096 45 0 0 L. 0 94607604219838 17364433920 7564054593336869
87 1 0 7 4096 45 0 0 L. 0 94607604219838 17364433920 7564054593336869
0 451 0 31 4096 63 0 0 L. 0 94602019621982 17364433920 0
451 0 451 31 4096 63 0 0 L. 0 94602019621982 17364433920 0
451 1 0 31 4096 63 0 0 L. 0 94602019621982 17364433920 0
0 86 0 31 4096 72 0 0 L. 0 94609539513109
...

Re: High Load Issues [message #12658 is a reply to message #12633] Sun, 06 May 2007 16:58 Go to previous messageGo to next message
dev is currently offline  dev
Messages: 1693
Registered: September 2005
Location: Moscow
Senior Member

ok. you don't have cpulimited VEs.
No you need to dump:
# ps axf -o pid,wchan,comm
to see which processes are in D state and where.


http://static.openvz.org/userbars/openvz-developer.png
Re: High Load Issues [message #12660 is a reply to message #12658] Sun, 06 May 2007 18:07 Go to previous message
rickb is currently offline  rickb
Messages: 368
Registered: October 2006
Senior Member
I think its worth noting--

As Dev pointed out, you do not have any cpulimits set on your VEs. This is generally bad especially if each VE has its own admin; ie you want to ensure fariness rather then optimal overall efficiency. When you add cpulimits to your VEs, the load will not decrease, and in fact it may increase. This is because a few of your VEs will reach their cpulimit and their runqueues will expand, which means more processes want to run as seen by your HN.

However, the end result is your cpuunits will kick into gear and each environment will have its amount of cpuunits cputime/equivalent MHz to count on. Each of your VEs will feel more interactive when cpuunits are well enforced, and cpulimit will protect the HN from reaching 0% idle time very often, which makes things even more interactive.

Be sure that your cpuunits are not oversubscribed. See my message #8620 on this page to read more about cpuunits/cpulimit and how they relate to each other. http://forum.openvz.org/index.php?&t=msg&th=1551

The bottom line is that the load as seen by the HN is not always an accurate factor in determining if your server is overcommitted, underpowered, etc (you may have plenty of idle cpu time with a load of 100!). With regard to strictly your application load, it may and probably is overcommited; however if you are not the admin of your VEs, like in the case you are selling them as a service, you are fulfilling your responsibility in guaranteeing them a time slice of the cpu / equivalent CPU speed in MHz and the HN load is not something you can control.

Hope this isn't too wordy!



-------------
Common Terms I post with: http://wiki.openvz.org/Category:Definitions

UBC. Learn it, love it, live it: http://wiki.openvz.org/Proc/user_beancounters
Previous Topic: *CLOSED* "TOP" inside a VPS
Next Topic: *CLOSED* Memory
Goto Forum:
  


Current Time: Wed Feb 21 02:55:35 GMT 2024

Total time taken to generate the page: 0.03071 seconds