# Resource usage for chef server

**URL:** <https://discourse.chef.io/t/resource-usage-for-chef-server/1221>\
**Category:** Chef Infra (archive)\
**Created:** [July 3, 2009, 5:20am UTC](https://discourse.chef.io/t/resource-usage-for-chef-server/1221 "2009-07-03T05:20:08Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Joe\_Van\_Dyk](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.chef.io/joe_van_dyk/32/853_2.png) [@Joe\_Van\_Dyk](https://discourse.chef.io/u/Joe_Van_Dyk)\
**Post date:** [July 3, 2009, 5:20am UTC](https://discourse.chef.io/t/resource-usage-for-chef-server/1221/1 "2009-07-03T05:20:08Z")

</div>

Should I be expected to run chef on a 256 megabyte server, if that’s all  
it’s doing?

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND

23665 couchdb 20 0 340m 40m 1300 S 0 15.9 38:16.26 beam.smp

23881 chef 20 0 111m 38m 1788 S 0 14.9 2:04.86 ruby

4848 root 20 0 86952 26m 1776 S 0 10.4 28:59.37 chef-client

3615 chef 20 0 86340 23m 1192 S 0 9.2 54:01.36 chef-indexer

23884 chef 20 0 109m 14m 1684 S 0 5.8 0:10.02 ruby

3617 root 20 0 69164 9852 992 S 0 3.8 17:08.68 stompserver

7272 www-data 20 0 726m 5632 2056 S 0 2.1 19:06.03 apache2

7373 www-data 20 0 726m 5592 2052 S 0 2.1 19:03.68 apache2

It’s swapping a bit.

---

<div class="post-metadata">

**Author:** ![Adam\_Jacob](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.chef.io/adam_jacob/32/292_2.png) [@Adam\_Jacob](https://discourse.chef.io/u/Adam_Jacob)\
**Post date:** [July 6, 2009, 5:34pm UTC](https://discourse.chef.io/t/resource-usage-for-chef-server/1221/2 "2009-07-06T17:34:17Z")

</div>

It's going to depend on what you are serving, and how. If you are  
really memory constrained, I would start looking at running a pair of  
thins and chef-client via cron.

Adam

On Thu, Jul 2, 2009 at 10:20 PM, Joe Van Dyk[joe@fixieconsulting.com](mailto:joe@fixieconsulting.com) wrote:

> Should I be expected to run chef on a 256 megabyte server, if that's all  
> it's doing?
> 
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> 
> 23665 couchdb 20 0 340m 40m 1300 S 0 15.9 38:16.26 beam.smp
> 
> 23881 chef 20 0 111m 38m 1788 S 0 14.9 2:04.86 ruby
> 
> 4848 root 20 0 86952 26m 1776 S 0 10.4 28:59.37 chef-client
> 
> 3615 chef 20 0 86340 23m 1192 S 0 9.2 54:01.36 chef-indexer
> 
> 23884 chef 20 0 109m 14m 1684 S 0 5.8 0:10.02 ruby
> 
> 3617 root 20 0 69164 9852 992 S 0 3.8 17:08.68 stompserver
> 
> 7272 www-data 20 0 726m 5632 2056 S 0 2.1 19:06.03 apache2
> 
> 7373 www-data 20 0 726m 5592 2052 S 0 2.1 19:03.68 apache2
> 
> It's swapping a bit.

--  
Opscode, Inc.  
Adam Jacob, CTO  
T: (206) 508-7449 E: [adam@opscode.com](mailto:adam@opscode.com)

---

<div class="post-metadata">

**Author:** ![Arjuna\_Christensen](https://avatars.discourse-cdn.com/v4/letter/a/5f9b8f/32.png) [@Arjuna\_Christensen](https://discourse.chef.io/u/Arjuna_Christensen)\
**Post date:** [July 6, 2009, 5:37pm UTC](https://discourse.chef.io/t/resource-usage-for-chef-server/1221/3 "2009-07-06T17:37:38Z")

</div>

On 6/07/2009, at 10:34 AM, Adam Jacob wrote:

> It's going to depend on what you are serving, and how. If you are  
> really memory constrained, I would start looking at running a pair of  
> thins and chef-client via cron.

Might I recommend a lighter weight webserver, Nginx for example?

> > 7272 www-data 20 0 726m 5632 2056 S 0 2.1 19:06.03 apache2
> > 
> > 7373 www-data 20 0 726m 5592 2052 S 0 2.1 19:03.68 apache2

> Adam
> 
> On Thu, Jul 2, 2009 at 10:20 PM, Joe Van  
> Dyk[joe@fixieconsulting.com](mailto:joe@fixieconsulting.com) wrote:
> 
> > Should I be expected to run chef on a 256 megabyte server, if  
> > that's all  
> > it's doing?
> > 
> > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> > 
> > 23665 couchdb 20 0 340m 40m 1300 S 0 15.9 38:16.26 beam.smp
> > 
> > 23881 chef 20 0 111m 38m 1788 S 0 14.9 2:04.86 ruby
> > 
> > 4848 root 20 0 86952 26m 1776 S 0 10.4 28:59.37 chef-  
> > client
> > 
> > 3615 chef 20 0 86340 23m 1192 S 0 9.2 54:01.36 chef-  
> > indexer
> > 
> > 23884 chef 20 0 109m 14m 1684 S 0 5.8 0:10.02 ruby
> > 
> > 3617 root 20 0 69164 9852 992 S 0 3.8 17:08.68  
> > stompserver
> > 
> > It's swapping a bit.
> 
> --  
> Opscode, Inc.  
> Adam Jacob, CTO  
> T: (206) 508-7449 E: [adam@opscode.com](mailto:adam@opscode.com)

--  
AJ Christensen, Software Engineer  
Opscode, Inc.  
E: [aj@opscode.com](mailto:aj@opscode.com)

---

<div class="post-metadata">

**Author:** ![Arjuna\_Christensen](https://avatars.discourse-cdn.com/v4/letter/a/5f9b8f/32.png) [@Arjuna\_Christensen](https://discourse.chef.io/u/Arjuna_Christensen)\
**Post date:** [July 6, 2009, 5:46pm UTC](https://discourse.chef.io/t/resource-usage-for-chef-server/1221/4 "2009-07-06T17:46:58Z")

</div>

On 6/07/2009, at 10:34 AM, Adam Jacob wrote:

> It's going to depend on what you are serving, and how. If you are  
> really memory constrained, I would start looking at running a pair of  
> thins and chef-client via cron.

Blaming the web server was a little out of line of me, and my knee  
jerk reaction is to install a lighter weight reverse proxy; however;  
depending on the number of nodes you're using you may not need to run  
a reverse proxying webserver at all (apache, nginx, lighttpd,  
haproxy, .., ..) unless you require SSL connection termination.

Hope this helps!

> On Thu, Jul 2, 2009 at 10:20 PM, Joe Van  
> Dyk[joe@fixieconsulting.com](mailto:joe@fixieconsulting.com) wrote:
> 
> > Should I be expected to run chef on a 256 megabyte server, if  
> > that's all  
> > it's doing?
> > 
> > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> > 
> > 23665 couchdb 20 0 340m 40m 1300 S 0 15.9 38:16.26 beam.smp
> > 
> > 23881 chef 20 0 111m 38m 1788 S 0 14.9 2:04.86 ruby
> > 
> > 4848 root 20 0 86952 26m 1776 S 0 10.4 28:59.37 chef-  
> > client
> > 
> > 3615 chef 20 0 86340 23m 1192 S 0 9.2 54:01.36 chef-  
> > indexer
> > 
> > 23884 chef 20 0 109m 14m 1684 S 0 5.8 0:10.02 ruby
> > 
> > 3617 root 20 0 69164 9852 992 S 0 3.8 17:08.68  
> > stompserver
> > 
> > 7272 www-data 20 0 726m 5632 2056 S 0 2.1 19:06.03 apache2
> > 
> > 7373 www-data 20 0 726m 5592 2052 S 0 2.1 19:03.68 apache2
> > 
> > It's swapping a bit.
> 
> --  
> Opscode, Inc.  
> Adam Jacob, CTO  
> T: (206) 508-7449 E: [adam@opscode.com](mailto:adam@opscode.com)

--  
AJ Christensen, Software Engineer  
Opscode, Inc.  
E: [aj@opscode.com](mailto:aj@opscode.com)

---

<div class="post-metadata">

**Author:** ![Albert\_Llop](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.chef.io/albert_llop/32/885_2.png) [@Albert\_Llop](https://discourse.chef.io/u/Albert_Llop)\
**Post date:** [July 6, 2009, 9:01pm UTC](https://discourse.chef.io/t/resource-usage-for-chef-server/1221/5 "2009-07-06T21:01:08Z")

</div>

our experience is that a 256mb slice is slightly small for the resources  
needed for a chef-server. You need two "instances" of the server working,  
and that's as many rails apps (i guess merb requires a little less) as you  
can get to work "properly" in a 256mb slice, but with chef-server you also  
need couchdb and stompserver. For us it didn't work properly until we  
upgraded to a 512mb slice.

my suggestion is to use a 360mb slice from linode (for example, costs the  
same as a 256mb slicehost), or to upgrade your current 256 to a higher plan,  
and use the extra resources for other things you might need. Remember that  
with a 256mb slice you're "losing" a good chunk on basic system processes,  
whereas when you upgrade to a higher plan, all that extra memory is "clean",  
as opposed to having two 256mb slices.

## Also nginx is pretty light, yeah 🙂

Albert Llop  
Tech Department

> **[Kings of Mambo](https://kingsofmambo.com/)**
>
> Creative and graphic studio from Barcelona. We make digital products strongly focused on the best UX/UI possible. We are also specialists in modern front-end and back-end development with Wordpress, October CMS & Yii.

2009/7/6 Arjuna Christensen [aj@opscode.com](mailto:aj@opscode.com)

> On 6/07/2009, at 10:34 AM, Adam Jacob wrote:
> 
> It's going to depend on what you are serving, and how. If you are  
> really memory constrained, I would start looking at running a pair of  
> thins and chef-client via cron.
> 
> Blaming the web server was a little out of line of me, and my knee jerk  
> reaction is to install a lighter weight reverse proxy; however; depending on  
> the number of nodes you're using you may not need to run a reverse proxying  
> webserver at all (apache, nginx, lighttpd, haproxy, .., ..) unless you  
> require SSL connection termination.
> 
> Hope this helps!
> 
> On Thu, Jul 2, 2009 at 10:20 PM, Joe Van Dyk[joe@fixieconsulting.com](mailto:joe@fixieconsulting.com)  
> wrote:
> 
> Should I be expected to run chef on a 256 megabyte server, if that's all
> 
> it's doing?
> 
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> 
> 23665 couchdb 20 0 340m 40m 1300 S 0 15.9 38:16.26 beam.smp
> 
> 23881 chef 20 0 111m 38m 1788 S 0 14.9 2:04.86 ruby
> 
> 4848 root 20 0 86952 26m 1776 S 0 10.4 28:59.37 chef-client
> 
> 3615 chef 20 0 86340 23m 1192 S 0 9.2 54:01.36 chef-indexer
> 
> 23884 chef 20 0 109m 14m 1684 S 0 5.8 0:10.02 ruby
> 
> 3617 root 20 0 69164 9852 992 S 0 3.8 17:08.68 stompserver
> 
> 7272 www-data 20 0 726m 5632 2056 S 0 2.1 19:06.03 apache2
> 
> 7373 www-data 20 0 726m 5592 2052 S 0 2.1 19:03.68 apache2
> 
> It's swapping a bit.
> 
> --  
> Opscode, Inc.  
> Adam Jacob, CTO  
> T: (206) 508-7449 E: [adam@opscode.com](mailto:adam@opscode.com)
> 
> --  
> AJ Christensen, Software Engineer  
> Opscode, Inc.  
> E: [aj@opscode.com](mailto:aj@opscode.com) [adam@opscode.com](mailto:adam@opscode.com)