When my OpenVZ VPS is running with lighttpd, RAM usage is very low, but when I switch to Apache, RAM usage goes all the way up to the max. I've had this happen to two VPSs already and can't figure out why this happens.
I also have two other VPSs, one Xen and one Virtuozzo and have no troubles with Apache.
How I can see how much guaranteed memory I have ? And how I can see how much guaranteed memory I have used ?
Additional how I can see how much bustable memory I have used or how much total memory my VPS is using.
Let say your VPS has 128MB of memory. My understanding would be that if you had, say, another 128MB of 'burstable' memory, that if another guest on the system wasn't using all it's memory, or if there was some spare on the system, then your VPS could use that.
But, how much memory does your VPS really think it has available? and what happens if the other VPS (or the system) decides it wants it memory back?
Does OpenVZ just start killing off your processes until it free's up enough memory?
How to limit memory for VE on openvz
Boys how to limit the memory Guaranteed of one VE in MB example:
128MB
256MB
512MB
1024MB
2048MB
How to set this to the VE?
Few months ago I bought new small VPS box (OpenVZ, 128 MB RAM) in order to place there a new monitoring node of my site monitoring system. Such small amount of RAM is a challenge for operating system optimisation techniques (OpenVZ doesn’t have “swap” as Xen does).
First of all I discovered that apache2-mpm-worker (Apache implementation that uses threads) consumes more memory (100MB) than the classic version that use separate processes (20MB). I had to switch to apache2-mpm-prefork version then.
Next unpleasant suprise: small Python app eats 100MB of virtual memory! I checked that virtual (not resident) memory is taken into account by VPS. I applied some tools to locate memory bottleneck, but without success. Next I added logs with current memory usage to track call that causes big memory consumption. I tracked the following line:
server = WSGIServer(app)
is guilty for high memory increase. After few minutes of googling I located problem: default stack size for a thread. Details:This line creates few threads to handle concurrent calls
Stack size is counted towards virtual memory
Default stack size is very high on Linux (8MB)
Every thread uses separate stack
=> multi threaded application will use at least number_of_threads * 8MB virtual memory!
First solution: use limits.conf file. I altered /etc/security/limits.conf file and changed default stack size. But I couldn’t make this change to alter Python scripts called from Apache (any suggestions why?).
Second (working) solution: lower default stack size using ulimit. For processes launched from Apache I altered /etc/init.d/apache2 script and added:
ulimit -s 256
Now every thread (in apache / Python application) will use only 128 kB of virtual memory (I lowered VSZ from 70 MB to 17 MB this way). Now I have additional space to enlarge MySQL buffers to make DB operations faster.
There’s even better place to inject ulimit system-wide: you can insert this call in:
/etc/init.d/rc
script. Then ulimit will be applied to all daemons (as Apache) and all login sessions. I reduced virtual memory usage by 50% this way.
Note: you may increase stack size on stack overflow errors. In my opinion 256 kb is safe option for most systems, you may increase if in doubt. Still memory savings are big.
We're in the process of setting up our new VPS Server, and we can create a VPS with 256MB memory and with 512MB memory fine, but when creating one with 1GB memory, we get the error:
Could Not Start Vps, Reason: Unable to fork: Cannot allocate memory: Not enough resources to start environment: Container start failed:
Even though the server has 4GB RAM and no other VPS's running. Any ideas?
Thanks.
[Edit]We now seem to get the problem for all our VPS's. I think it may be something to do with the Server not unallocating the memory, as we've provisioned and de-provisioned quite a few Servers
[openvz ]How to make for kernel of openvz to recognize 4GB of ram?
How to make for kernel of openvz to recognize 4GB of ram?
# uname -a
Linux xxxxxxxxxxx 2.6.18-92.1.13.el5.028stab059.3 #1 SMP Wed Oct 15 17:48:55 MSD 2008 i686 i686 i386 GNU/Linux
I am using VPS (Linux OS), and web server is apache. I totally have 768M memory. When the system start up, I usually have 550M-600M free memory.
I just found that each httpd process occupied almost 20M memory. It's horrible, I was though that 1M for each process is reasonable.
Now the httpd.conf looks like:
StartServers 5
MinSpareServers 5
MaxSpareServers 10
MaxClients 150
But if I change StartServers to 50, httpd will eat up all my memory. I did some calculation, that 600M/20M = 30. This means if apache have more than 30 concurrent requests, There server will be out of memory. The "MaxClients 150" is meaningless! I was though 150 is too small.
I only use apache with php, they are complied by WHM. Should I re-complied the apache with php?
Configuration: Xeon 3060 4gb of ram
So mysql is using 6% of memory, xcache is set to 256MB, and rest is for apache.
So my question is what do I need to do? Compile apache to have smaller memory foot print or cut down more quicker idlers? Usually there is no problem, but if I got like 300+ people browsing the joomla site all day eventually system gets out of memory and starts swaping. Any ideas?
Quote:
Server Version: Apache/1.3.37 (Unix) mod_gzip/1.3.26.1a mod_auth_passthrough/1.8 mod_log_bytes/1.2 mod_bwlimited/1.4 PHP/4.4.7 FrontPage/5.0.2.2635.SR1.2 mod_ssl/2.8.28 OpenSSL/0.9.7a
Server Built: May 31 2007 00:49:37
Quote:
Current Time: Monday, 24-Sep-2007 19:54:50 CEST
Restart Time: Monday, 24-Sep-2007 12:19:45 CEST
Parent Server Generation: 27
Server uptime: 7 hours 35 minutes 5 seconds
Total accesses: 691782 - Total Traffic: 2.7 GB
CPU Usage: u199.12 s35.93 cu.09 cs.71 - .864% CPU load
25.3 requests/sec - 102.3 kB/second - 4135 B/request
30 requests currently being processed, 31 idle servers
Quote:
total used free shared buffers cached
Mem: 4054 3227 826 0 44 456
-/+ buffers/cache: 2727 1326
Swap: 1992 125 1867
Quote:
netstat -an|awk '/tcp/ {print $6}'|sort|uniq -c
27 ESTABLISHED
6 FIN_WAIT1
11 FIN_WAIT2
3 LAST_ACK
39 LISTEN
11 SYN_RECV
1 SYN_SENT
204 TIME_WAIT
Quote:
Timeout 45
KeepAlive On
MaxKeepAliveRequests 300
KeepAliveTimeout 2
MinSpareServers 10
MaxSpareServers 20
StartServers 10
MaxClients 300
MaxRequestsPerChild 1000
Code:
nobody 24280 0.2 4.3 418304 179380 ? S 17:01 0:24 /usr/local/apache
mysql 4071 4.9 6.1 290892 255556 ? S<l Jul23 4523:49 /usr/sbin/mysqld
nobody 24812 0.2 4.2 418180 176320 ? S 17:08 0:25 /usr/local/apache
nobody 25029 0.2 1.2 296428 52092 ? S 17:11 0:24 /usr/local/apache
nobody 25107 0.2 3.5 417672 147076 ? S 17:12 0:21 /usr/local/apache
nobody 26743 0.2 4.1 418276 173520 ? S 17:34 0:20 /usr/local/apache
nobody 31438 0.2 3.4 417864 142332 ? S 18:37 0:10 /usr/local/apache
nobody 31443 0.3 4.0 417720 169660 ? S 18:37 0:13 /usr/local/apache
nobody 31444 0.2 4.0 417112 167412 ? S 18:37 0:12 /usr/local/apache
nobody 31445 0.2 4.1 417820 171336 ? S 18:37 0:12 /usr/local/apache
nobody 31446 0.2 1.0 294716 45512 ? S 18:37 0:10 /usr/local/apache
nobody 31447 0.2 4.1 417668 171076 ? S 18:37 0:12 /usr/local/apache
nobody 32514 0.2 3.2 416880 135924 ? S 18:52 0:07 /usr/local/apache
root 32578 0.0 0.0 7456 2340 ? Ss 18:53 0:00 sshd: root@pts/2
I got apache 1.3.3.7 installed and i really do not know what to do. I though I already optimized server but after 6 days of uptime memory is full and i just saw load went to 3 and started to swap. Specs are RHEL 4, xeon 3060 with 4gb of ram.
Settings:
Timeout 300
KeepAlive On
MaxKeepAliveRequests 200
KeepAliveTimeout 4
MinSpareServers 10
MaxSpareServers 15
StartServers 10
MaxClients 200
I'm under impression that if i add 64gb of ram i would still fill it up in lets say 15 days because apache isn't releasing memory properly. After load 3 some memory is released:
Mem: 4054 3171 882 0 22 259
-/+ buffers/cache: 2890 1163
Swap: 1992 330 1662
I got atm 220 users online and when i restart apache even if i have 500 concurrent users online usage of memory is only 800mb, and as u see now its 3gb and only half of users are online so its not really peak time.
Apache processes are still hanging from 4 hours ago and they are not cut, as you can see every process takes around 4% of memory:
nobody 19618 0.1 4.1 281596 172400 ? S 18:36 0:20 /usr/local/apache/bin/httpd -DSSL
nobody 21100 0.2 4.4 288872 183672 ? S 18:57 0:22 /usr/local/apache/bin/httpd -DSSL
nobody 22654 0.2 1.5 161620 66168 ? S 19:18 0:22 /usr/local/apache/bin/httpd -DSSL
nobody 23250 0.2 1.4 161740 60564 ? S 19:28 0:20 /usr/local/apache/bin/httpd -DSSL
nobody 23423 0.2 4.2 283636 176940 ? S 19:30 0:22 /usr/local/apache/bin/httpd -DSSL
nobody 23479 0.1 2.3 283652 96900 ? S 19:31 0:16 /usr/local/apache/bin/httpd -DSSL
So i really hope there is solution because restarting apache every few days isn't fun and it sounds insane.
Apache is wasting away a bit over a GB of memory at about 100 connections. Is this something normal? Or should I be receiving a little better performance?
Also, I find that the load is very low. Is there anyway to sort of trade a bit of the CPU Usage for some more memory?
I have a VPS with 256MB of guaranteed memory, and no burst/swap. Having little memory like this is quite common for a VPS and I want to know how to make the most of it while running websites. By far the biggest culprit for taking my memory is Apache and I'm not convinced that the default setup makes sense on a VPS.
I know there are other daemons out there that use considerably less memory, such as lighttpd, but enough people (myself included) are going to use Apache regardless (they might require some of the advanced features after all) so it is worth knowing how to get the most out of it.
Apache 2 has various mpm options, most notably prefork and worker. As I understand it the prefork method forks processes to handle requests ahead of time, and can fork more on demand if required, each process handles one request at a time. Worker uses threads, again created ahead of time and on demand, each thread handles one request at a time.
With the default configuration and prefork Apache spawned 6 processes each taking about 5MB, so there goes 50MB in total. With workers it wanted to create 50 threads, each with a 10MB stack, this instantly went over the memory limit so wouldn't start at all. Thankfully the stack can be adjusted using the ThreadStackSize setting so it can be made to run, but even after this what do you really have for your memory?
One problem with having one request per process/thread is that, again by default, Apache has a keep alive setting of 15 seconds. This means that once a request has been processed, the process/thread is then kept around for up to 15 seconds in case the client wants to do anything else. With prefork that is 5MB of memory being used, and even with threads with a reasonable 1MB stack size there is still a huge amount of memory wastage on a connection that might not even do anything. Meanwhile other visitors might be getting out-of-memory problems as Apache tries to spawn more handlers to meet demand.
I think it should be fairly obvious that the worker mpm is the best option here, if everything works fine with a 1MB stack then the thread overhead is considerably lower than the process overhead.
The new events mpm should relieve this problem as it uses an event queue to keep an eye on keep alive connections rather than a whole thread. Unfortunately this mpm is not considered stable yet, so in the mean time I am wondering if VPS users should just disable keep alive, or at least adjust it to just a few seconds.
The benefit of keep alive is that it reduces network traffic and CPU load as there is a small amount of overhead in setting up a new TCP connection. When servers had 400MHz chips and massive 256MB of ram this was a good trade off. But now VPS often have access to much more powerful processors, and the same 256MB isn't considered massive any more so a lot of applications are less frugal and less adept at running in this environment.
In the unlikely even that Apache had 200MB to play with, using 1MB threads means that the system can handle 200 idle connections before running out of resources and failing to handle new connections. With a keep alive of 15 seconds you can support 200 connections per 15 seconds. In light of the fact that many newer browsers will open 4 simultaneous connections (up from the previous 2) that is only 3.34 users per second. Drop the keep alive to 1 second and you can handle 50 users per second. These numbers are unrealistic because they assume the request handling itself takes no time at all, but a lot of requests can be handled in split seconds so maybe these numbers aren't so far fetched. If the request handling takes 250ms and you turn keep alive off you could handle 200 users per second.
I am thinking that it would be great to have a reverse proxy sat in front of Apache, one that used an event queue and thus had little memory overhead for maintaining incoming connections, but closed down its connection to Apache after each request. Unfortunately I do not know of one.
my grunted memory for apatchi is 51 meg
i want make him 256meg
because the vps its alwayes down
see this please
HTML Code:
29890 nobody 0
2.5
0.1 /usr/local/apache/bin/httpd -k start -DSSL
31854 nobody 0
2.3
0.1 /usr/local/apache/bin/httpd -k start -DSSL
31766 nobody 0
1.8
0.1 /usr/local/apache/bin/httpd -k start -DSSL
32043 nobody 0
1.7
0.1 /usr/local/apache/bin/httpd -k start -DSSL
25921 nobody 0
1.4
0.1 /usr/local/apache/bin/httpd -k start -DSSL
32014 nobody 0
1.3
0.1 /usr/local/apache/bin/httpd -k start -DSSL
25998 nobody 0
1.3
0.1 /usr/local/apache/bin/httpd -k start -DSSL
30388 nobody 0
1.0
0.1 /usr/local/apache/bin/httpd -k start -DSSL
Which is more efficient in the way of memory? Using Subversion with Apache and WebDav, or using SVNServe? I figure the latter, but by how much? How much ram usually would be used by the former? I'm undecided as to which method I want to use to set up Subversion yet.
View 3 Replies View RelatedI currently have 2 VPSs that I'm monitoring and I noticed that on one of them the memory usage is much higher than the other which doesn't make sense to me. I tracked the high memory usage to the Apache processes.
Here's the overview of the VPSs:
LiquidWeb VPS- 512MB RAM, 10 Sites with minimal usage.
Average httpd process (Owner Priority Cpu% Mem% Command):
nobody 0 0.0 1.2 /usr/local/apache/bin/httpd -k start -DSSL
Virpus VPS- 384MB RAM, 61 sites with moderate usage.
Average httpd process (Owner Priority Cpu% Mem% Command):
nobody 0 0.0 0.3 /usr/local/apache/bin/httpd -k start -DSSL
Both Apache configurations are identical, the only difference besides the hardware is that LiquidWeb runs CentOS4.7 and Virpus runs CentOS5.2, any ideas? It's not to big of a deal since I'm not near my limit for either of them but I was just curious.
I've got a complicated problem and I can't figure out what I need to edit to fix this.
I've been messing with the httpd.conf and php.ini files trying to optimize them as best I can. I've managed to get the load to stay pretty low all day, and everything seemed to be working fine. But then apache crashes.
(See screenshot for details)
httpd starts out with 70-150 VIRT memory, and it constantly goes up all day, and I've yet to see it go down. And, eventually it just crashes Apache! In the screenshot it's 502m, but I've seen that get as high as 800! And sometimes I even see a number like this "80376" no "m" or anything, just "80376"; I've no idea what that's about.
So what do I need to edit or tweak to fix this?
I'm on a cPanel linux server, Apache 2.2.8 and PHP 5.2.6.
tell me what the average usage of the CPU and RAM be from apache?
User 'Nobody' (Which I presume is apache) is using 13% CPU and 38.98% of RAM. Is this normal or is it well overloaded?
In the last day or two I've been having issues with memory on one of by boxes.
Something eats it all up, so the OS starts swapping, the I/O wait shoots up, and soon the load is up in the hundreds and the thing is totally useless.
During the day today I've tracked the something to occational apache processes. It seems that occationally a thread is started which uses upwards of 150M of memory. These threads are obvioudly doing something heavyweight and take a while to complete. When I get a few of them running together it soons finishes off all my available memory. Below is an extract from top when a couple of these threads are running.
Code:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
10230 nobody 25 0 197m 27m 9392 R 45.5 2.9 0:03.27 /usr/local/apache/bin/httpd -DSSL
10231 nobody 25 0 197m 25m 8376 R 52.8 2.7 0:01.60 /usr/local/apache/bin/httpd -DSSL
I was wondering if anyone could help out by suggesting some ways of tracking down what page/script it which uses so much memory. It's a cPanel server so it's not really practical to tail -f the apache logs (not knowing which account it is means I don't know which file to watch).
I'm running a single Magento website with only around 150 products, and very low traffic, on a VPS with 2 Cores and 4GB of ram. However I am getting constant warning emails about high Apache Memory usage and High CPU usage.
The server and site were working fine, however since the server failed (Power cut on upgrade at the hosts end) and we rebuilt the website I'm getting these warnings, my host said it is just the website and the server needed upgrading... to which I did (server was a 1 core 2GB ram server). However since I've upgraded I'm still getting these warning emails.
Here is the latest one: URL...
Code:
4951 apache 20 0 441m 137m 6556 S 0.0 3.6 2:28.77 httpd
4952 apache 20 0 443m 139m 7168 S 0.0 3.6 2:43.72 httpd
7377 apache 20 0 358m 54m 6576 S 0.0 1.4 1:25.76 httpd
7692 apache 20 0 433m 129m 6588 S 0.0 3.4 1:20.64 httpd
[code]....
I'm really new to VPS's and plesk so I'm wondering what I can do to try and track down what is causing these high usage as I'm 100% certain the site was fine before. The server has default build both before and after the fail.
Server Information:
OSCentOS 6.5 (Final)
Panel version11.5.30 Update #47
2 Cores
4GB Ram
Are there any tools within plesk that I can use to see what is causing these issues? It's making the website load extremely slow.
I'm having an issue with what I believe is Apache. Things will be running nicely, then all of a sudden things will just 'blow up'. Apache seems to fork a bunch of new processes, and it starts ripping into the memory. I've had it essentially eat through all the free memory and swap, and the system will then freeze up as there's no more memory. This is a CentOS server running CPanel, the configuration had been up and running for a long time without problems then this suddenly started happening.
View 7 Replies View RelatedWe are currently running plesk 11.5 and over the last few months, the apache memory usage has been very high. On investigation we have gone through the logs and we can't find a cause.
I can issue a service httpd restart and the memory drops for about 30 minutes and then we see the apache memory grow to almost 1.5 - 2GB. Why this keeps happening?
I have a VPS. And have had an issue both when it was 1Gig and now I recently downgraded it to 768m, because I am moving some sites to a dedicated.
However, the part I am having trouble grasping is that when I look at graphs from Munin, it will typically always show 200-400MB free memory (and free -m and top agrees with munin), but Munin shows 'committed' memory that is above the total Ram on the VPS and once the 'committed' ram exceeds the VPS limit, processes start failing.
So, why is 'committed' memory exceeding the RAM on my VPS, when Munin, free -m and top all show there is free memory available?
Code:
root@server [~]# free -m
total used free shared buffers cached
Mem: 768 449 318 0 0 0
-/+ buffers/cache: 449 318
Swap: 0 0 0
Here's a graph that munin produces that shows the 'committed' memory exceeding the total memory.
[url]
I just got a new server Dual E5520 with 6GB RAM, SAS 15k rpm raid10. It's running well. However, the memory usage is just around 2.5GB, even when I have more traffic. Here is the kernel info
Quote:
# uname -a
Linux server2.[url]2.6.18-128.1.10.el5 #1 SMP Thu May 7 10:35:59 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
Any idea that we can put more content into memory?
I just ordered a server from serverbeach that should have come with 2 GB but I see that this one comes with 8 GB
Unfortunately, I seem to be missing 7 GB of ram.
Installed Physical Memory: 8 GB
Total Physical Memory: 1 GB
Windows 2008 Server Web 32-bit
Now, being a 32-bit system, I should see 3-4 GB of ram right?
Any idea why only 1 GB is available for the OS?
[url]
I have a 512mb DV server with Mediatemple, which I am running 24 (ish) domains off (most of them static websites) and a teamspeak server. I would say MAX theres 10 users online at a time)
Now, I know its running out of memory because i get frequent QoS Alerts in plesk (kmemsize is apparently the memory size):
Oct 01, 2009 11:52:57 AMBlack zonekmemsize
I have attached my results (when I did top).
My questions are:
1. Should I be expecting to be out of memory running what I am?
2. Is there a way to see the problem domains (memory wise)?
3. Are there any ways I can reduce the memory? (I have followed this already:
[url]
4. Where is the memory usage coming from (I am finding it very hard to understand TOP)
Can anybody advise, or provide somewhere which advises on the pros/cons of OpenVZ and Xen virtualisation platforms?
View 14 Replies View RelatedWell which one?
Also if possible say why you prefer either OpenVZ or Xen.
I am being told that when a VPS in openvz hits the maximum ram assigned to it that openvz disables that vps.
Isn't the ram limit a hard limit that simply restricts the available ram to that container to XX ram?
Is this normal?
We are looking for a VPS to house a scala framework running on a JVM, from what I can gleen Xen would be the best route for us or is it possible that openVZ with enough dedicated ram would suffice, does anyone have a view on this?
View 6 Replies View Related