CentOS 5 :: Determine What Is Causing High Load
Sep 12, 2011
I am running a CentOS 5.6 32 bit installation under VMWare ESXi and have been experiencing some very high load values from time to time. The server is running multiple installations of gameservers, and the load fluctuates from around 1 to 9, with more or less the same amount of players playing (give or take 5-10 of a possible 80). I've been running DStat with almost all possible metrics, and, except for the large fluctuations in load, nothing else out of the ordinary seems to happen when load is rising. (Log data can be provided, if anyone wants to see it). Disk IO, Network throughput, memory consumption, CPU usage, process count, all stay at the same levels when the load is 1 as when it is 9. How I can begin to troubleshoot this, and find out why the load goes to such high values?
View 1 Replies
ADVERTISEMENT
Apr 5, 2009
I recently installed Centos-5.3-x86_64 in a PowerEdge2950 machine, to serves as an ftp server using proftpd-1.3.2. I noticed that the CPU load is much higher than the days when I was using RHEL-4.7-X86_64. In the past, CPU load get higher than 100 only when there're about 400 online users. But in Centos-5.3, the CPU load gets higher than 150 when there're 200 users.
The details of the server machine: Intel Xeon E5430, 16 GB RAM, 1.8 TB as a RAID disk, 400GB hardisk used to installed the system.
More details:
View 11 Replies
View Related
Nov 24, 2010
I have several CentOS 32-bit VMs running on ESX 4. Those that were updated to the most current patch level ("yum update", accepting all updates available last week) started showing load average of ~0.4 when completely idle. After comparing the problematic VMs with those that show zero load average at idle, and then modifying them in all kinds of ways, I narrowed it down to the combination of a recent kernel patch (2.6.18-164 is fine, 2.6.18-194 is not) and E1000 network adapter. Replacing the network adapter on a problematic VM with VMXNET3 fixes the load statistics.
View 3 Replies
View Related
Dec 29, 2010
Just trying to figure out some stuff with a broken process. A java app seems to sometime get stuck on a loop or something and i'm trying to find out what's causing it using just the following #sysadmin tools at my disposal.
Things like:- htop - find the PID thats causing the High CPU cycles. I'd then want to use /proc/[PID] or lsof -p [PID] or strace [PID] etc. But the PID doesn't exist in 'ps -ef' output so I think htop must be showing me kernel level thread PIDs?
Not sure about the PID's HTOP is actually giving me? I know that some of them are the real PID's that can be accessed through /proc/[pid] etc but others are not but are i assume child processes or more likely threads as child processes are normally shown in a default ps output anyway.
Is someone able to help distinguish to me about what exactly all these other PID's are that I can't manipulate or find apart from when using HTOP.
View 7 Replies
View Related
May 11, 2011
After GRUB 2 comes up (I'm running Ubuntu 10.10) and I choose the OS to boot, there is about a 5 second delay where nothing appears to happen after I make the selection -- no disk activity. It happens consistently every time I boot. Again, this is after I choose the OS to boot, so it shouldn't have anything to do with the standard delay to allow me to choose the appropriate OS.Is there a good way to troubleshoot this and determine what is causing the delay?
View 4 Replies
View Related
Feb 10, 2010
my server and it had a load of 60 so i immediately took down apache and the load went back down to 0 in a few mins and every time i put it back on the cpu usage on both cores immediately goes to 100% and the load goes up to 20 in just 60 seconds until i take apache down again?
View 6 Replies
View Related
Aug 14, 2010
Something is running my memory up and causing my load to crash my forum. I keep noticing this error and have no clue really how to go about fixing it.
Quote:
%CPU PID USER COMMAND
1.2 2978 mysql /usr/sbin/mysqld --basedir=/ --datadir=/var/lib/mysql --user=mysql --log-error=/var/lib/mysql/server.forumnameremoved.com.err --pid-file=/var/lib/mysql/server.forumnameremoved.com.pid
Could this be causing my CPU to run high? How would I go about fixing an error like this or is it even doable?
View 9 Replies
View Related
May 2, 2011
Whenever someone copies a data Our system get very slow. Load average of system starts increasing. How would i find which process is causing the same. I have tried following top 10 CPU utilizing process command but it didn't help much.
Code:
ps -auxf | sort -nr -k 3 | head -10
My distribution detail.
Code:
# lsb_release -a
LSB Version: :core-3.1-ia32:core-3.1-noarch:graphics-3.1-ia32:graphics-3.1-noarch
Distributor ID: EnterpriseEnterpriseServer
[code]....
View 4 Replies
View Related
Sep 15, 2015
I have a debian system with the following version you can se bellow. My problem is that one one single core of 4 it's running 100% all the time and i cant seem to find out why. The load is also high(load average: 0.91, 0.75, 0.40) because of this. This keeps happening even after reboots. The system if freshly installed twice and the same problem occours as soon as it boots. Something called kworker is running and causing the load / CPU load, is that normal for a fresh install??
root@Cyberdyne:/# lsb_release -da
No LSB modules are available.
Distributor ID: Debian
Description: Debian GNU/Linux 8.2 (jessie)
Release: 8.2
Codename: jessie
View 5 Replies
View Related
Aug 10, 2011
I wrote a script to extract and get the the name of *.gz in a foler . Since running that script every 10 minutes, load average on my server increases more than 10.I checked with 'top' and it showed many D process.
View 2 Replies
View Related
Apr 2, 2011
on an old server of mine, as soon as apache is started, the load average that I get to see with 'top', that normally is under 1, now just steadily climbs up and up to easily 150, in fact disabling the webserver from serving any webpage. I've checked netstat, and I'll try to upload the output. The ip's that are in there I've blocked with iptables. But that doesn't help or so it seems. I see nothing weird in the error logs. As soon as I stop apache, the load goes back to normal. As soon as I (re)start it, up it goes again. What can cause this and how do I get rid of it?
p.s. It's an old server, fedora3 or so, and I've got a new one to which I'll transfer the domains, but until that's completely done, I'd like this one to run as it has for years...
View 1 Replies
View Related
Dec 14, 2009
The load average is almost 1.06, but cpu is not 100% utilized... I am just wondering why load average is 1.06 in that case still ? (well i monitored cpu for long time, it never exceeded above 40%)can anyone explain the reason behind it ? Also is the system over utilized in this case ?
View 3 Replies
View Related
Aug 18, 2011
I am having Red Hat LINUX 5 Enterprise Server and facing problem regarding very very high server load (load average is going high up to 60-70)due to which server is getting hang.
View 7 Replies
View Related
Jul 17, 2009
I was wondering how can I determine among the modules loaded at boot which of them are really necessary and which are not, in order to reduce the boot process time and have a more "elegant" system start.
I know this theme is a little bit of complicated because it depends of the user's point of view and demand a high knowledge of which things are happening in your system but I need somewhere to start improving the performance of my debian system.
View 3 Replies
View Related
May 5, 2010
Since upgrading from 9.10 to 10.04, my system is noticeably less responsive and exhibits halting behavior for at times 10s of seconds. There is nothing obvious in the various logs that I can find. The only objective indication I can find that demonstrates that something is seriously wrong is that my load average never drops below 0.5 and is often over 1.5, even at idle. In this situation, top/ps/whatever shows very few processes running (usually just top). This suggests to me that either the new kernel scheduler is horrible or that something new is resulting in blocking I/O or other uninterruptable sleeps.Typical top output:
Quote:
top - 12:18:01 up 1 day, 49 min, 5 users, load average: 1.42, 0.86, 0.71
Tasks: 262 total, 1 running, 261 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.9%us, 0.8%sy, 0.0%ni, 97.6%id, 0.0%wa, 0.0%hi, 0.7%si, 0.0%st
[code]....
View 9 Replies
View Related
Jul 2, 2010
i've been using 10.04 since may and it's been good to me thus far. BUT recently, i've been experiencing unusually high system loads. it will happen at random it seems. for instance, it just did it a few minutes ago when i turned on and mounted a external firewire drive. i also had ogmrip open, but NOT running as well as transmission and pidgin. yesterday, i was encoding a rather big movie with ogmrip and around 25%, the system load quickly increased to 10 and held there until i force quit ogmrip. it went back to normal after the force quit.
a previous night, i was installing the popular ubuntu themes via the terminal. at the same time, i was also using ogmrip and the system load shot up during the download/install and held there the rest of the night. sooo...unless it's somehow related to ogmrip i don't know what's going on. it even did it when ogmrip WAS NOT processing any video, just simply open.
View 2 Replies
View Related
Jul 6, 2010
I am facing a high load averages in one of my linux server find the output of the top command.
[code]...
How to reduce the load averages.
View 3 Replies
View Related
Nov 14, 2010
During downloads, the top command shows the Firefox process at 100% CPU. Yesterday I tried to download an .iso image. After a few hours the Firefox window would not refresh nor would it respond to input. I tried the wget command. It used negligible CPU time and completed in 28 minutes.
This problem is easy to reproduce because it happens every time I download a file in Firefox. It also happens when I use a fresh profile to run Firefox without any extensions or plugins.
This is Firefox 3.6.12 on Fedora 14.
View 3 Replies
View Related
Jul 28, 2010
I have a apache cluster with more than 10 nodes, based on ldirector and heartbeat. The problem is that I cannot predict if my nodes will handle the traffic in the next day (hosting a website based on daily campaigns). So I decided to limit the number of active connection on the nodes (from apache), but this is only a temporary solution. I want to create a page that will appear to users that are getting over the limit. Has anyone made this before? Can you tell me how is it possible (I don't want a how to, just a starting point to study)? I think squid can do it, but I don't know how to search for it. To give you an example of what I want, you can see the same thing on deviantart.
View 3 Replies
View Related
Feb 10, 2010
I have High load on my server and my investigation shows nothing (so i believe that my investigations is wrong ), the load average in this moment is 10.13, 9.47, 8.24. , mentioning the below.
- The disk utilization (all the disks) is near 0, as the result of the IOSTAT
- There is no blocked processes (as a result of VMSTAT).
- I have two processors (dual core) , the maximum load average should be something around 4.
- The server always have above 8 load average in all times interval.
btw , my OS is RHEL AS release 4 (Nahant Update 7)Kernel :Linux 2.6.9-78.ELhugemem #1 SMP i686 i686 i386 GNU/Linux
View 8 Replies
View Related
Jan 13, 2010
I have high load average on the servers with no identified reason (From my point of view)
View 2 Replies
View Related
Jan 9, 2011
I just posted about this in this thread, but as the other thread was started by a KDE user then I thought I'd post here as well. I've had high CPU usage for a few months now - probably since trying the 0.9 branch of Compiz then dropping back to the default openSUSE builds (XOrg and gconfd-2 running a Core i5 at about 30% on every core*). I've now finally found a solution after deciding I wanted to fix it once and for all.
Once again, the Ubuntu forums come to the rescue with this thread (I don't like the distro as a whole, but I do find the forums useful!). I'm using Compiz, but it turns out that Metacity was running as well. A quick "killall -9 metacity" and the gconfd-2 process has vanished and XOrg settled down to its normal 1-2% (which is reasonable when I've got a Conky config refreshing every fraction of a second to repaint a sound visualiser!). Now I just need to find out why Metacity starts when I'm using Compiz...
* according to Conky's per-core graphs, although top only reported 15% overall and the Conky "top 3 procs by CPU" reported a measly 3% for each process, so someone's maths was out somewhere!
View 3 Replies
View Related
May 10, 2010
I upgraded webserver to new ubuntu server 10.04 (x86-64). After upgrade the increased load from 0,3 to 1,4. On webserver running phpbb, which generating slow quieres, which not before upgrade to lucid. HW conf: Intel Core i7, 8GB ram, WD Raptor 10k rpm. Week17 upgrade to new version.
View 9 Replies
View Related
Oct 15, 2010
It's the fourth time now since Maverick that I had to cold reboot my system because it was totally unresponsive. The system monitor in my taskbar shows 100% on the background blue graph (I guess that's just one core then) and almost nothing on the other core. On the last freeze I managed to open 'sudo top' before it went totally unresponsive and saw that there was no high CPU usage at all, but a load average spiking above 24. Also the swap seemed to be full although my machine usually never uses swap. I was watching a movie with VLC this time, but I'm not sure if VLC ran the other times my OS froze. I made a snapshot with my cell phone: [URL]. How I can prevent a process from causing so much load?
View 6 Replies
View Related
Nov 26, 2010
I am having a problem with the server that I use to host my personal site. The load average quite often spikes to exceed 1.00 for the 1 and 5 minute intervals, and the 15 minute interval gets above .5. This occurs while the server is idle, serving very few or no requests and with the CPU 99% idle with <1% IOWAIT usage. I have checked top and vmstat, but neither one provides any useful info. Top continues to say the CPU is 99% idle, and vmstat says that there are 0 runnable and 0 blocking tasks. Occasionally, vmstat will say that there is 1 runnable task, but this doesn't even coincide with the load average spikes. I have already searched for other solutions to this problem, but everything I have seen says to use top and/or vmstat, but those aren't showing anything out of the ordinary. Can anyone recommend anything else I might do?
My server has a Pentium 4 HT 3gHz processor, 2GB RAM, and runs Kubuntu 10.10. (The reason it runs Kubuntu instead of Ubuntu Server is that it needs an X environment so that the Nvidia driver can initialize and put its graphics card into a power saving mode.)
View 8 Replies
View Related
Jul 6, 2010
I have a Nagios server with a lot of hosts and services; around 400 services (in all) and 150 hosts. Most of these services are programmed in bash language. The problem is the server has a high load average; between 5 and 11. The server has the next features:
- Intel Xeon 2.66GHz Dual Core
- 4MB cache memory
- 1GB RAM memory
- 50GB hard disk
Is this load average normal? Should I program the plugins with C?
View 3 Replies
View Related
Jul 16, 2010
Just noticed that during calls skype uses ~30-50% cpu, + pulseaudio uses ~20%. I found some old threads on this like [URL] Or [URL]. The former suggests to purge pulse and use alsa/oss instead. The latter suggests changing mic in skype to DeviceXX and tweak the pulse.conf
There's also a launchpad bug on that with status Confirmed->invalid (due to "problem fixed in skype 2.1 beta") but looks like it's not (I am using 2.1 beta)
1. Does any1 has this problem?
2. I can't try 2nd link's approach as my skype has only PulseAudio in the settings, so I can't select Device or anything else!
3. Should I try to remove pulse in favor of ALSA? I'm pretty happy with pulse otherwise that this issue, I use it to record sound out of my sound card and I'm not sure if I can do this with alsa/os.
View 4 Replies
View Related
Oct 31, 2010
After changing servers and moving things around, I have one system which serves the /home directories, and for some reason, it is extremely slow. I've looked for duplicate IPs, signs of a HW problem, etc. and cannot figure out why the system is running with a high load, high iowaits, and low responsiveness. The servers and clients are slackware 12.2 and 13.1 respectively.
View 3 Replies
View Related
Apr 10, 2011
After installing reddit, my ubuntu stuck with high cpu usage and system load.
Code:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
8634 reddit 20 0 12180 8960 2572 R 12.1 0.9 0:00.51 paster
8611 root 20 0 23756 17m 3816 R 11.8 1.7 0:00.99 paster
8613 reddit 20 0 23200 17m 3908 R 11.8 1.7 0:00.96 paster
8617 reddit 20 0 15072 11m 2688 R 11.8 1.1 0:00.69 paster
8620 reddit 20 0 14372 10m 2680 R 11.8 1.1 0:00.66 paster
[Code]...
View 1 Replies
View Related
Sep 10, 2010
I've just noticed that "nicing" long running computationally intensive, I/O unintensive, single-threaded executables on my system increases the CPU run time of those executables (as reported by /usr/bin/time as well as by wall clock) by a factor of 2-3 even in the absence of any load, i.e., "top" tells me that my program is getting 100% of a CPU. For example:
%coot 248: /usr/bin/time -p ./a.out > out
real 97.47
user 85.64
[code]....
My system is running Ubuntu 7.10. If I run the same executable on two other machines I have access to -- one running Fedora 5 and the other Ubuntu 9.10 -- I don't see any discrepancy between the runtimes using nice and not using nice.
This behavior is executable independent, compiler dependent, and language dependent -- I'm seeing it across the board. I'm assuming I've somehow configured my system to behave this way, but I have no idea what I may have done. Also, this was the first time I'd ever done timing runs with "nice" (actually, "at"), so I'm not sure how long my system's been configured (if configured is the right word) this way.
View 4 Replies
View Related
