Re: [gitorious] Self-hosted - slow and unresponsive, taking 100% CPU time

2012-07-28 Thread Donald Henderson
Thank you - that was the issue - switching to synchronous messaging helped 
but it wasn't enough - so more RAM it was 
All the best
Donald

On Tuesday, July 17, 2012 7:32:35 AM UTC+1, Marius Mårnes Mathiesen wrote:

 On Mon, Jul 16, 2012 at 11:21 PM, Donald Henderson  wrote:

 I have just installed my own deployment of Gitorious. It's a Debian 
 Squeeze box, hosted by Linode, 512MB of RAM and shared access to a 
 quad-core 3GHz Xeon. Running Ruby 1.8 and the latest rails. Only change I 
 had to make from the standard procedure was update Rots manually as bundle 
 wasn't pulling it from GitHub. 


 I'm afraid 512 megs of RAM isn't going to get you very far :-( The 
 Gitorious Rails environment will take a few hundred megs, and you'll need 
 at least two instances to get any work done (a webapp instance and a 
 message queue poller).  

 Each time your server goes from serving you web pages to doing background 
 work, it will have to do a lot of swapping; this could explain why the app 
 feels reasonably snappy until you upload an SSH key (which will be 
 processed by the message queue poller).

 One thing you could try is to switch to syncronous messaging, this will 
 save you the memory required for the message queue poller - but you really 
 should get some more memory.

 Cheers,
 - Marius

 The problem I'm having is that it's slow to the point of being unusable. 
 There is nothing else running on the server except exim. I've tried from 
 both Chrome and Firefox and am seeing the issue with both, so I don't think 
 it's the browser. 

 When I first hit the homepage, whether already logged in or not, it takes 
 a genuine 5 seconds to load, with a 99-100% ruby1.8 process running. 
 Browsing speeds up after that, but when I try to, for example, add an SSH 
 key or configure teams, it hangs and doesn't respond. I have restarted the 
 server (and apache2) and these have reset the connection. I have also left 
 it for up to 5 minutes before giving up. 

 There is nothing good in log/production.log but a lot of 
 connect failed: Connection refused - connect(2) will retry in 5
 in the apache error log 

 I have read 
 https://groups.google.com/forum/#!msg/gitorious/2Y6wwE_V7Vc/nIfeAiW0IkQJ and 
 the stomp server seems to be installed and running 

 Can anyone point me to what might be wrong? 

  -- 
 To post to this group, send email to gitorious@googlegroups.com
 To unsubscribe from this group, send email to
 gitorious+unsubscr...@googlegroups.com




 -- 
 Marius Mårnes Mathiesen
 Rubyist, Shortcut AS
 Tel.: (+47) 92 60 95 38.

 http://shortcut.no


-- 
To post to this group, send email to gitorious@googlegroups.com
To unsubscribe from this group, send email to
gitorious+unsubscr...@googlegroups.com


Re: [gitorious] Self-hosted - slow and unresponsive, taking 100% CPU time

2012-07-17 Thread Marius Mårnes Mathiesen
On Mon, Jul 16, 2012 at 11:21 PM, Donald Henderson 
donald.henderso...@gmail.com wrote:

 I have just installed my own deployment of Gitorious. It's a Debian
 Squeeze box, hosted by Linode, 512MB of RAM and shared access to a
 quad-core 3GHz Xeon. Running Ruby 1.8 and the latest rails. Only change I
 had to make from the standard procedure was update Rots manually as bundle
 wasn't pulling it from GitHub.


I'm afraid 512 megs of RAM isn't going to get you very far :-( The
Gitorious Rails environment will take a few hundred megs, and you'll need
at least two instances to get any work done (a webapp instance and a
message queue poller).

Each time your server goes from serving you web pages to doing background
work, it will have to do a lot of swapping; this could explain why the app
feels reasonably snappy until you upload an SSH key (which will be
processed by the message queue poller).

One thing you could try is to switch to syncronous messaging, this will
save you the memory required for the message queue poller - but you really
should get some more memory.

Cheers,
- Marius

The problem I'm having is that it's slow to the point of being unusable.
 There is nothing else running on the server except exim. I've tried from
 both Chrome and Firefox and am seeing the issue with both, so I don't think
 it's the browser.

 When I first hit the homepage, whether already logged in or not, it takes
 a genuine 5 seconds to load, with a 99-100% ruby1.8 process running.
 Browsing speeds up after that, but when I try to, for example, add an SSH
 key or configure teams, it hangs and doesn't respond. I have restarted the
 server (and apache2) and these have reset the connection. I have also left
 it for up to 5 minutes before giving up.

 There is nothing good in log/production.log but a lot of
 connect failed: Connection refused - connect(2) will retry in 5
 in the apache error log

 I have read
 https://groups.google.com/forum/#!msg/gitorious/2Y6wwE_V7Vc/nIfeAiW0IkQJ and
 the stomp server seems to be installed and running

 Can anyone point me to what might be wrong?

  --
 To post to this group, send email to gitorious@googlegroups.com
 To unsubscribe from this group, send email to
 gitorious+unsubscr...@googlegroups.com




-- 
Marius Mårnes Mathiesen
Rubyist, Shortcut AS
Tel.: (+47) 92 60 95 38.

http://shortcut.no

-- 
To post to this group, send email to gitorious@googlegroups.com
To unsubscribe from this group, send email to
gitorious+unsubscr...@googlegroups.com