Hi
I have a centos instance on a host where uname -a shows:
Linux cl28810.com 2.6.18-194.26.1.el5.028stab079.2PAE #1 SMP Fri Dec 17 19:34:22 MSK 2010 i686 i686 i386 GNU/Linux
Periodically my ssh access and http access slow dramatically and I get out of memory when trying to do rudimentary stuff like ls. I've been advised by tech support to try service httpd restart. This works. But the first thing that happens after the command is issued is that the load goes up to about 15 then slowly declines to a more typical figure, in the range feom .1 to 1.3, which is fine. It takes several minutes before the site is really usable. I now have 22 apache instances out of 66 total processes, which seems excessive.
One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, Xeon based host. I don't know if this is an apache issue or CentOS related or what.
Ideas? Debugging I can do?
Dave
centos-bounces@centos.org wrote:
One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, Xeon based host. I don't know if this is an apache issue or CentOS related or what.
My first recommendation: You could double (or quadruple) that RAM far cheaper than talking about it.
Insert spiffy .sig here: Life is complex: it has both real and imaginary parts.
//me ******************************************************************* This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. www.Hubbell.com - Hubbell Incorporated**
2011/3/17 Brunner, Brian T. BBrunner@gai-tronics.com:
centos-bounces@centos.org wrote:
One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, Xeon based host. I don't know if this is an apache issue or CentOS related or what.
My first recommendation: You could double (or quadruple) that RAM far cheaper than talking about it.
is this openvz based virtual machine?
-- Eero
On Thursday, March 17, 2011 01:09:46 pm Eero Volotinen wrote:
2011/3/17 Brunner, Brian T. BBrunner@gai-tronics.com:
centos-bounces@centos.org wrote:
One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, Xeon based host. I don't know if this is an apache issue or CentOS related or what.
My first recommendation: You could double (or quadruple) that RAM far cheaper than talking about it.
is this openvz based virtual machine?
virtuozzo is what the hosting org says
d
-- Eero _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Dave Stevens wrote:
Periodically my ssh access and http access slow dramatically and I get out of memory when trying to do rudimentary stuff like ls. I've been advised by
<snip>
is fine. It takes several minutes before the site is really usable. I now have 22 apache instances out of 66 total processes, which seems excessive.
One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, Xeon based host. I don't know if this is an apache issue or CentOS related or what.
Ideas? Debugging I can do?
For one thing, I'd strongly urge you to add at least another .5G RAM, if not 2.5G. Second, look at the apache configuration, and see how many workers it can use - you might want to lower the limit.
Also, what's your connection to the 'Net like, and how heavily is your website being hit?
When facism comes to America it will be wrapped in the flag and carrying a cross.
- Sinclair Lewis, It Can't Happen Here, 1935
It's happened here.
mark
2011/3/17 m.roth@5-cent.us:
Dave Stevens wrote:
Periodically my ssh access and http access slow dramatically and I get out of memory when trying to do rudimentary stuff like ls. I've been advised by
<snip> > is fine. It takes several minutes before the site is really usable. I now > have 22 apache instances out of 66 total processes, which seems excessive. > > One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, > Xeon based host. I don't know if this is an apache issue or CentOS related > or what. > > Ideas? Debugging I can do?
For one thing, I'd strongly urge you to add at least another .5G RAM, if not 2.5G. Second, look at the apache configuration, and see how many workers it can use - you might want to lower the limit.
Also, what's your connection to the 'Net like, and how heavily is your website being hit?
machine looks openvz virtualized container, it requires tuning to limit settings.
-- Eero
On Thursday, March 17, 2011 01:09:48 pm m.roth@5-cent.us wrote:
Dave Stevens wrote:
Periodically my ssh access and http access slow dramatically and I get out of memory when trying to do rudimentary stuff like ls. I've been advised by
<snip>
is fine. It takes several minutes before the site is really usable. I now have 22 apache instances out of 66 total processes, which seems excessive.
One site hosted, running drupal 7, 1.5G RAM lots of disk spece, two cores, Xeon based host. I don't know if this is an apache issue or CentOS related or what.
Ideas? Debugging I can do?
For one thing, I'd strongly urge you to add at least another .5G RAM, if not 2.5G. Second, look at the apache configuration, and see how many workers it can use - you might want to lower the limit.
ok, I'll look at that
Also, what's your connection to the 'Net like,my measured download to the
server is 40 mbps
and how heavily is your website being hit?
almost not at all, not in production
When facism comes to America it will be wrapped in the flag and carrying a cross.
- Sinclair Lewis, It Can't Happen Here, 1935
It's happened here.
mark
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 3/17/2011 4:55 PM, Dave Stevens wrote:
Ideas? Debugging I can do?
For one thing, I'd strongly urge you to add at least another .5G RAM, if not 2.5G. Second, look at the apache configuration, and see how many workers it can use - you might want to lower the limit.
ok, I'll look at that
This is probably really a drupal issue. I don't use it, but I've seen comments about it using a lot of memory and you can probably find tips on how to tune it. The multiple instances of apache should mostly share memory if the startup is smart enough to preload common data before forking.
This is not CentOS, it's a compilation by your host built from changed CentOS sources. Your problem obviously is that you reach the limits of your VM. 1.5 GB of RAM should be enough for your 22 apache processes, so it is probably the CPU that is the bottleneck. Anyway, you have to talk to your provider, this is nothing CentOS-specific.
Kai
Greetings,
On 3/18/11, Dave Stevens geek@uniserve.com wrote:
Hi
I have a centos instance on a host where uname -a shows:
Linux cl28810.com 2.6.18-194.26.1.el5.028stab079.2PAE #1 SMP Fri Dec 17 19:34:22 MSK 2010 i686 i686 i386 GNU/Linux
I would suggest you look into the swap settings on both the Centos* host and the Centos* guest.
Is swap stored on RAID0+1 etc.
Is the VM stored in LVM or on filesystem, if any, which?
It is also difficult to make out if the 1.5G RAM you mentioned is physically on the host or allocated to VM.
Having some such details would make an interest furtherance to this thread.
HTH
Regards,
Rajagopal
On Thursday, March 17, 2011 02:31:28 pm Rajagopal Swaminathan wrote:
Greetings,
On 3/18/11, Dave Stevens geek@uniserve.com wrote:
Hi
I have a centos instance on a host where uname -a shows:
Linux cl28810.com 2.6.18-194.26.1.el5.028stab079.2PAE #1 SMP Fri Dec 17 19:34:22 MSK 2010 i686 i686 i386 GNU/Linux
I would suggest you look into the swap settings on both the Centos* host and the Centos* guest.
Is swap stored on RAID0+1 etc.
Is the VM stored in LVM or on filesystem, if any, which?
It is also difficult to make out if the 1.5G RAM you mentioned is physically on the host or allocated to VM.
Having some such details would make an interest furtherance to this thread.
HTH
Regards,
Rajagopal
Thanks to all who replied, I'm taking the matter forward with the hosting outfit.
Dave
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Dave Stevens wrote:
Rajagopal Swaminathan wrote:
Greetings,
Dave Stevens geek@uniserve.com wrote:
Hi
I have a centos instance on a host where uname -a shows:
Linux cl28810.com 2.6.18-194.26.1.el5.028stab079.2PAE #1 SMP Fri Dec 17 19:34:22 MSK 2010 i686 i686 i386 GNU/Linux
I apologize for this late reply, have been occupied elsewhere.
Consider configuring Apache's server-status feature: http://httpd.apache.org/docs/2.2/mod/mod_status.html It won't specifically help with memory issues, at least not directly, but it is a good starting point when tuning Apache.
Here are specific Apache tuning tips: http://httpd.apache.org/docs/2.2/misc/perf-tuning.html
You can place a test-load on your server and get some baseline statistics about the server's performance characteristics using Apache's benchmark utility: /usr/sbin/ab -n 1000 -c 10 http://localhost/server-status so you'll have a point of reference for where your attempts to tune it are going.
I wouldn't expect RAM to be the limiting factor for Apache, perhaps some feature in Drupal is leaking memory? What does top say about the
-
Brunner, Brian T. -
Charles Polisher -
Dave Stevens -
Eero Volotinen -
Kai Schaetzl -
Les Mikesell -
m.roth＠5-cent.us -
Rajagopal Swaminathan