Guest xaav Posted August 18, 2011 Posted August 18, 2011 I looked at htop when this was happening, and it appears you are putting too much pressure on MySQL.
Ice IT Support Posted August 20, 2011 Author Posted August 20, 2011 How can I be putting pressure on mySQL when I'm not on the site when this is happening?
Krydos Posted August 20, 2011 Posted August 20, 2011 I don't think xaav was refering to you personally when he used the word you, and furthermore this thread is discussing why Helionet is unstable. Helionet is these forums and the main website, or also known as cody; there are no user accounts on cody. While simultaneously watching htop running on cody and johnny I can say that there are plenty of times when cody will be running really sluggish and laggy but only using ~5% cpu. Meanwhile johnny will be using the other 95% plus more if he could get it and the physical processors are pegged out at 100%. I don't know how the virtualization is set up, but is there anyway to reserve some cpu time specifically for cody to use? Perhaps there is some way to set all the johnny processes a lower priority and the cody processes a higher cpu priority with nice? I have seen mysql hogging a lot of processor, but then again I have seen cron, ror, mongrel, etc really hog all the cpu too. It's definitely johnny that is causing the problems though. Hopefully some of these ideas and observations might help.
Guest xaav Posted August 21, 2011 Posted August 21, 2011 I think it's gotten better after I turned on caching, however what would really be great would be for regulation on IO from vmware.
jje Posted August 25, 2011 Posted August 25, 2011 This is interesting... in the last 12 hours Johnny has been down but Cody has been flawless:
Recommended Posts