burgemaster Posted July 11, 2011 Posted July 11, 2011 (edited) We are soon (hopefully!) launching a wordpress based intranet site that is hosted on one of our local servers. The server is a HP Proliant DL380, 4GB Ram with 3 x 72GB SCSI Drives, these are in RAID5 formatted with one simple volume. It was setup with Ubuntu and currently using port:80 It works perfectly when one user is accessing the site but when 5 or more users access the first default page it grinds to a halt. Can anyone please point in the direction that we should try and follow to find out what could be causing this? Is the server maybe not upto spec? We plan to have 60+ users logging in at once. Thanks in advance Edited July 11, 2011 by burgemaster
morganw Posted July 11, 2011 Posted July 11, 2011 Run top and iotop and see which processes are causing the problem. If the server looks okay then check you have a proxy exception so that you are accessing it over the LAN not through your internet connection.
CyberNerd Posted July 11, 2011 Posted July 11, 2011 The server should easily cope with that load. I'm not sure what the current state of php accelerators is with ubuntu (IIRC newer versions of php have an accelerator by default?), but after checking out morganw's advice I would look into php accelerators and php configs. some tweaks here (i've not used) 13 Great WordPress Speed Tips & Tricks for MAX Performance - Noupe Design Blog
mattian Posted July 11, 2011 Posted July 11, 2011 Run top and iotop and see which processes are causing the problem. If the server looks okay then check you have a proxy exception so that you are accessing it over the LAN not through your internet connection. Hi Have run top, here is the output when simulating 20 users wcess. org.uk/Capture1 . JPG Remove space etc as it's my first post.
morganw Posted July 11, 2011 Posted July 11, 2011 Are those spiked values or is the CPU constantly being used like that? It looks like a problem with Apache. Is this server just based on choosing a LAMP installation from tasksel or have you set everything up separately?
mattian Posted July 12, 2011 Posted July 12, 2011 Spiked, as soon as I set the simulation going it creaps up and stays at that level untill I stop it. If I run it for a while I get the following: [Mon Jul 11 15:38:15 2011] [error] child process 1785 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1786 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1787 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1788 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1789 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:16 2011] [notice] caught SIGTERM, shutting down This is a default LAMP install - the only thing I have done is enable mod_rewrite Can post anything you want that may help?
mattian Posted July 12, 2011 Posted July 12, 2011 Interestingly - this is idle... Tasks: 241 total, 1 running, 240 sleeping, 0 stopped, 0 zombie Cpu(s): 0.1%us, 0.1%sy, 0.0%ni, 99.8%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 3990672k total, 3941460k used, 49212k free, 15588k buffers Swap: 5804028k total, 2970388k used, 2833640k free, 145636k cached is that not saying the memory is totaly in use even when idle?
mattian Posted July 12, 2011 Posted July 12, 2011 Latest, Natty. Original install was 10.10 then I upgraded it the other day. Its the command line server edition
morganw Posted July 12, 2011 Posted July 12, 2011 I would reinstall a clean LAMP server into a VM and see if you get the same issues. There is a bug report for similiar issues but that is from a while ago. If the VM doens't work then I would suspect that there is something up with the Wordpress install that is causing problems, if you snapshot the VM before you install Wordpress you could also test install similar system like Joomla or Moodle and see if you are getting the same problem. I've not used Wordpress before but it might be worth checking permissions on correct on the install, you could chmod 777 the entire install just for testing purposes and see what happens.
mattian Posted July 12, 2011 Posted July 12, 2011 I have done a fresh wordpress install on a new database. This is what I have got a result of: top - 10:45:49 up 20:29, 3 users, load average: 13.04, 7.88, 3.62 Tasks: 109 total, 11 running, 98 sleeping, 0 stopped, 0 zombie Cpu(s): 91.0%us, 4.5%sy, 0.0%ni, 3.7%id, 0.0%wa, 0.0%hi, 0.8%si, 0.0%st Mem: 3990672k total, 1048072k used, 2942600k free, 14100k buffers Swap: 5804028k total, 34040k used, 5769988k free, 125204k cached It now seems to be using less memory, but the same CPU
dhicks Posted July 12, 2011 Posted July 12, 2011 [Mon Jul 11 15:38:15 2011] [error] child process 1785 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1786 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1787 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1788 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:15 2011] [error] child process 1789 still did not exit, sending a SIGKILL [Mon Jul 11 15:38:16 2011] [notice] caught SIGTERM, shutting down Apache pre-branches handler processes, ready to pass off the work for each request. You can set how many processes it pre-branches - I think the default could well be 5. It looks like all your handler processes are busy, which maybe implies a long-running or infinite loop in some bit of code that gets called when you visit the default page - probably a bit of PHP code somewhere.
morganw Posted July 12, 2011 Posted July 12, 2011 Wordpress recently (couple of months back) had malicious code inserted into their plug-ins, is this a brand new Wordpress install or are you moving it from somewhere else?
jamesfed Posted July 12, 2011 Posted July 12, 2011 (edited) Not that its a solution to your problem (not even a work around) but have you considered getting an install of Windows Server on your DL380 and run WordPress off of IIS? All our sites here are hosted on IIS7.5 (a few internal WordPress installs and HAP+) and it works a real treat. Not that I'm looking to bring up a discussion about this but it could be the easy way out if you have a lic spare... Edited July 12, 2011 by jamesfed
localzuk Posted July 12, 2011 Posted July 12, 2011 My first thought would be this - don't use a non LTS release of Ubuntu to run a server - there's no need, and it is more work than you want to do maintaining it. Instead, use the latest LTS release and go from there. They are more stable, tested more and supported for longer. However, in this case, with massive number of Apache processes being spawned for so little load, I'd be thinking something is wrong with Wordpress. Can I ask what you are using to simulate your load? Also, do you have anything like custom themes or plugins installed in Wordpress? (I'm thinking not as you mention a fresh install). Finally, how are you installing it? Custom site definitions or straight into the default /var/www/ location?
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now