Load Averages
list Gavin Leonard
I am getting pages from my hobbit server when some of my linux systems are hitting 15 on their load avg. does that equate to something worse than it sounds? Cause 15% load average on a linux system does not seem that bad? Or is it not a percentage? Gavin Leonard [cid:image001.gif at 01C92958.D1EDC010] Director, Systems-Network Engineering T XXX-XXX-XXXX F XXX-XXX-XXXX E user-d65663809eb4@xymon.invalid<mailto:user-d65663809eb4@xymon.invalid> Research | Marketing | Sales Generation www.progrexion.com<http://www.progrexion.com/>; This email and its contents are confidential. If you are not the intended recipient, delete this email and do not use or disclose the information within this email or its attachments. Thank you.
list Michael McCullen
Load is a measure of how many processes are running or runnable and is not directly related to % of CPU used. There are cases where Load average is through the roof but CPU % is very low, usually when when the server is having issues talking to NFS storage or accessing some kind of I/O resource. We generally monitor load average based on number of CPUs in the box. Our standard is yellow at 1.5 * # CPUs and red at 2 * # CPUs. YMMV.
▸
On Oct 8, 2008, at 5:16 PM, Gavin Leonard wrote:
I am getting pages from my hobbit server when some of my linux systems are hitting 15 on their load avg. does that equate to something worse than it sounds? Cause 15% load average on a linux system does not seem that bad? Or is it not a percentage? Gavin Leonard
<image001.gif>
Director, Systems-Network Engineering
T
XXX-XXX-XXXX
F
XXX-XXX-XXXX
E
user-d65663809eb4@xymon.invalid
Research | Marketing | Sales Generation
www.progrexion.com
▸
This email and its contents are confidential. If you are not the intended recipient, delete this email and do not use or disclose the information within this email or its attachments. Thank you.
list Rob MacGregor
▸
On Wed, Oct 8, 2008 at 22:16, Gavin Leonard <user-d65663809eb4@xymon.invalid> wrote:
I am getting pages from my hobbit server when some of my linux systems are hitting 15 on their load avg. does that equate to something worse than it sounds? Cause 15% load average on a linux system does not seem that bad? Or is it not a percentage?
http://en.wikipedia.org/wiki/Load_(computing) The Load Average is the number of processes requiring processor time. What it means depends on the number of processors - a load average of 6 is fine for a system with 8 processors, but not good news for one with only a single processor. Fifteen on an 8 processor box isn't good news, but isn't the end of the world. On a single processor box it's a sign that you've got problems ;) -- Please keep list traffic on the list. Rob MacGregor Whoever fights monsters should see to it that in the process he doesn't become a monster. Friedrich Nietzsche
list Josh Luthman
It is not a percentage. http://www.luv.asn.au/overheads/NJG_LUV_2002/luvSlides.html Simply put the load is how much calculator power your box is doing. I have seen servers that have a value of several hundred load on a day to day basis. The biggest reason for it being monitored is to watch what is usual and when you see a spike, ascend, descend, etc, you have a history to compare it to. This value will be different for nearly every single server. Josh Luthman Office: XXX-XXX-XXXX Direct: XXX-XXX-XXXX XXXX Wayne St Suite XXXX Troy, OH XXXXX Those who don't understand UNIX are condemned to reinvent it, poorly. --- Henry Spencer
▸
On Wed, Oct 8, 2008 at 5:16 PM, Gavin Leonard <user-d65663809eb4@xymon.invalid>wrote:
I am getting pages from my hobbit server when some of my linux systems are hitting 15 on their load avg. does that equate to something worse than it sounds? Cause 15% load average on a linux system does not seem that bad? Or is it not a percentage? *Gavin Leonard***
[image: cid:image001.gif at 01C856AD.922EF120]
▸
Director, Systems-Network Engineering
*T*
XXX-XXX-XXXX
*F*
XXX-XXX-XXXX
*E*
user-d65663809eb4@xymon.invalid
Research | Marketing | Sales Generation
*www.progrexion.com* <http://www.progrexion.com/>;
▸
This email and its contents are confidential. If you are not the intended
recipient, delete this email and do not use or disclose the information
within this email or its attachments. Thank you.
list Rich Smrcina
▸
Gavin Leonard wrote:
I am getting pages from my hobbit server when some of my linux systems are hitting 15 on their load avg. does that equate to something worse than it sounds? Cause 15% load average on a linux system does not seem that bad? Or is it not a percentage?
Your thinking CPU Utilization. Load Average measures the number of running processes. -- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2009 - Orlando, FL - May 15-19, 2009
list Greg L Hubbard
Gavin, Load average is (more or less) the average length of the run queue for a UNIX-style system (someone once posted a link to a more precise definition on this forum). For your system, a load average value of 15 means that an average of 15 processes are waiting to use the CPU -- a problem for most systems. If you look at the trends button, you will probably find a CPU utilization graph that shows a split between user, system, and idle CPU usage. CPU utilization shows how busy the CPU(s) is/are, but load average shows the impact of a busy CPU on the system's workload. I once had a system that spiked to 200+ in the load average display. This problem was traced to a problem in some multi-threaded code where the threads were going into infinite loops one by one until all of them were locked up. Amazingly enough, the system still responded to interactive usage, but v-e-r-y s-l-o-w-l-y. GLH
▸
From: Gavin Leonard [mailto:user-d65663809eb4@xymon.invalid]
Sent: Wednesday, October 08, 2008 4:16 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: [hobbit] Load Averages
I am getting pages from my hobbit server when some of my linux
systems are hitting 15 on their load avg. does that equate to something
worse than it sounds? Cause 15% load average on a linux system does not
seem that bad? Or is it not a percentage?
Gavin Leonard
Director, Systems-Network Engineering
T
XXX-XXX-XXXX
F
XXX-XXX-XXXX
E
user-d65663809eb4@xymon.invalid <mailto:user-d65663809eb4@xymon.invalid>
▸
Research | Marketing | Sales Generation www.progrexion.com <http://www.progrexion.com/>; This email and its contents are confidential. If you are not the intended recipient, delete this email and do not use or disclose the information within this email or its attachments. Thank you.
list Gavin Leonard
It is based on physical processors or cores? -Gavin
▸
From: Josh Luthman [mailto:user-4c45a83f15cb@xymon.invalid] Sent: Wednesday, October 08, 2008 3:28 PM To: user-ae9b8668bcde@xymon.invalid Subject: Re: [hobbit] Load Averages It is not a percentage. http://www.luv.asn.au/overheads/NJG_LUV_2002/luvSlides.html Simply put the load is how much calculator power your box is doing. I have seen servers that have a value of several hundred load on a day to day basis. The biggest reason for it being monitored is to watch what is usual and when you see a spike, ascend, descend, etc, you have a history to compare it to. This value will be different for nearly every single server. Josh Luthman Office: XXX-XXX-XXXX Direct: XXX-XXX-XXXX XXXX Wayne St Suite XXXX Troy, OH XXXXX Those who don't understand UNIX are condemned to reinvent it, poorly. --- Henry Spencer On Wed, Oct 8, 2008 at 5:16 PM, Gavin Leonard <user-d65663809eb4@xymon.invalid<mailto:user-d65663809eb4@xymon.invalid>> wrote: I am getting pages from my hobbit server when some of my linux systems are hitting 15 on their load avg. does that equate to something worse than it sounds? Cause 15% load average on a linux system does not seem that bad? Or is it not a percentage? Gavin Leonard
Error! Filename not specified.
▸
Director, Systems-Network Engineering T XXX-XXX-XXXX F XXX-XXX-XXXX E user-d65663809eb4@xymon.invalid<mailto:user-d65663809eb4@xymon.invalid> Research | Marketing | Sales Generation www.progrexion.com<http://www.progrexion.com/>; This email and its contents are confidential. If you are not the intended recipient, delete this email and do not use or disclose the information within this email or its attachments. Thank you.
list Hobbit User in Richmond
▸
On Wed, October 8, 2008 17:27, Josh Luthman wrote:
Simply put the load is how much calculator power your box is doing. I have seen servers that have a value of several hundred load on a day to day basis.
Well, it depends on what sort of work the box is doing, and what your throughput expectations are, but any server at a load average of 200+ is going to have abysmal response times for online users. I've seen that on dedicated mail servers in a spam storm, but when our Oracle database servers go above about 30, we start getting calls from application users. The default alert levels in Hobbit work pretty well for us, letting us check whether anything's going awry or if it's a true spike in demand, before the phones start ringing. I do keep the dedicated Internet-facing sendmail servers at significantly higher thresholds, because mail doesn't necessarily have to move in near real time, and customers are more tolerant of slow delivery than "unable to deliver" warnings and bounces.
list Rob MacGregor
▸
On Wed, Oct 8, 2008 at 22:38, Gavin Leonard <user-d65663809eb4@xymon.invalid> wrote:
It is based on physical processors or cores?
The number of CPU cores in total - so a pair of single core processors and one dual core processor are effectively the same.
▸
--
Please keep list traffic on the list.
Rob MacGregor
Whoever fights monsters should see to it that in the process he
doesn't become a monster. Friedrich Nietzsche