8 pointsGetting my WPMU DEV WingsI'm new here
gbrown88
Member
—
22nd March 2011 (1 year ago)
Hello,
I am having an issue with my current installation of WordPress MU (3.1) causing an EXTREMELY high query calls to the tmp place within the database. It is so high that is consistently crashing my host.
I am having trouble pinpointing the issue. I've tried shutting down many of the plugins, then doing a check and its still occurring.
I am wondering if someone has seen this before or they could suggest certain lines of code, files, etc to check for this.
I am having an issue with my current installation of WordPress MU (3.1) causing an EXTREMELY high query calls to the tmp place within the database. It is so high that is consistently crashing my host.
I am having trouble pinpointing the issue. I've tried shutting down many of the plugins, then doing a check and its still occurring.
I am wondering if someone has seen this before or they could suggest certain lines of code, files, etc to check for this.
We currently have 15 subdomains on our site. We are using the FeedWordPress plugin to feed in articles from 100+ sites to 5 of the subdomains. I've already turned this plugin off though, and the issue is still occurring.
We are also using:
WP-Related Posts
WP Super Cache
Dynamic Content Gallery
All in One SEO Pack
Embed Allower
Sitemaps and SEO - WordPress Multisite Style
Not sure about the "tmp" part but it could be similar (it seems so long ago I forget) to the issue I once had, I use to have a server hit all the time with SPLOGS, it would often bring it to its knees and nearly always around 2am - 4am (GMT). Our sys admins said the server was receiving high levels of connections which put huge strain on the resources often hitting pages/DB without referrals.
Is this your server?
What didn't help for us was back then is I had installed a good few plugins for use and they were not all optimal at what they do, you see both plugins and themes create entries in the database and when removing them they don't always remove the stuff they insert thus bloating your DB, it might not be to noticeable on one blog but in a Multisite install there are many.
Other things to note is sites running multiple plugins will make many calls and include many php files and depending on the purpose of the plugin can make many DB calls. Consider plugins which use the head hook for ads, social icons, widgets etc, etc.
We now make people apply for blogs through a contact form or our support desk and then create it for them. Automatic registration is off to avoid this issue unless you want to go into load balancing and such. In fact have you looked at the multi-DB plugin from WPMUDEV?
Have you also tried something like Doncha's Super Cache plugin?
If it is caused by surges in traffic could look at geo blocking IPs if the traffic causing this is coming from select countries or areas around the world. Most of ours was from China and Asia but keep in mind there are genuine users there and if this is your target market then it wouldn't be a viable option and it could cause issue for other genuine users. I'm not recommending this but it could be a consideration.
You haven't enabled post by e-mail or anything have you?
Anyway, this relates to my personal experience and could be different in your case.
We currently don't allow users to sign up for their own blogs either. We have been using WP Super Cache. Like you said, I think the related posts plugin may be the next one I stop and check to see if that helps.
It does sound like it may be a similar issue to yours though as are resources are going crazy at certain times.
If this is your server is it a cPanel/WHM server if so you could do your own checking and even terminate high processes as they occur. Those were not fun days. lol
You could look at configuring the CFS to be more strict and if you don't have the CSF firewall then consider getting it installed.
Is your smaller WPMU install more busy with traffic?
We just got a spike of traffic from another site so I asked them if thats the issue and this was their response:
That could certainly be possible but usually in situations like that the loud would be sustained for greater periods of time. What I'm seeing is a load spike out of relatively nowhere and then it settles back down again.
It does indeed sound like the issues we had, spikes which would amass rapidly then occasionally dissipate quickly or get much more intense quickly that you simply could not get into the server even over SSH. Ours was related to traffic, splogs and a bloated install.
Have you you checked the DB and ran any needed repairs, perhaps optimise the tables?
You could just try disabling every plugin for a few days and then turning just one back on every few days to give it time, maybe see if it is one of the plugins.
Just back to Supercache for moment have you considered:
Lock Down: Disabled
Prepare your server for an expected spike in traffic by enabling the lock down. When this is enabled, new comments on a post will not refresh the cached static files.
10635 pointsLike some sort of WPMU DEV GodMindblowingly helpful memberLifetime member
Sales & Support Lead
—
30th March 2011 (1 year ago)
#
Hiya gbrown88,
Hoping you got this one sorted as it's been awhile since you've posted. I'm marking this as resolved, but if you need further assistance feel free to re-open this thread.
Responses (14)
Member — 23rd March 2011 (1 year ago) #
More info....
We currently have 15 subdomains on our site. We are using the FeedWordPress plugin to feed in articles from 100+ sites to 5 of the subdomains. I've already turned this plugin off though, and the issue is still occurring.
We are also using:
WP-Related Posts
WP Super Cache
Dynamic Content Gallery
All in One SEO Pack
Embed Allower
Sitemaps and SEO - WordPress Multisite Style
Thanks
Grant
Support Chimp — 23rd March 2011 (1 year ago) #
Not sure about the "tmp" part but it could be similar (it seems so long ago I forget) to the issue I once had, I use to have a server hit all the time with SPLOGS, it would often bring it to its knees and nearly always around 2am - 4am (GMT). Our sys admins said the server was receiving high levels of connections which put huge strain on the resources often hitting pages/DB without referrals.
Is this your server?
What didn't help for us was back then is I had installed a good few plugins for use and they were not all optimal at what they do, you see both plugins and themes create entries in the database and when removing them they don't always remove the stuff they insert thus bloating your DB, it might not be to noticeable on one blog but in a Multisite install there are many.
Other things to note is sites running multiple plugins will make many calls and include many php files and depending on the purpose of the plugin can make many DB calls. Consider plugins which use the head hook for ads, social icons, widgets etc, etc.
We now make people apply for blogs through a contact form or our support desk and then create it for them. Automatic registration is off to avoid this issue unless you want to go into load balancing and such. In fact have you looked at the multi-DB plugin from WPMUDEV?
Have you also tried something like Doncha's Super Cache plugin?
http://wordpress.org/extend/plugins/wp-super-cache/
How aggressive would you be prepared to go?
If it is caused by surges in traffic could look at geo blocking IPs if the traffic causing this is coming from select countries or areas around the world. Most of ours was from China and Asia but keep in mind there are genuine users there and if this is your target market then it wouldn't be a viable option and it could cause issue for other genuine users. I'm not recommending this but it could be a consideration.
You haven't enabled post by e-mail or anything have you?
Anyway, this relates to my personal experience and could be different in your case.
Support Chimp — 23rd March 2011 (1 year ago) #
Ah just seen your second post - I was in the middle of writing mine as you posted it.
I'm not familiar with some of the plugins you mentioned. I should imagine related posts would make many DB queries though to math up those posts?
Support Chimp — 23rd March 2011 (1 year ago) #
Can your host not pinpoint what is escalating the DB calls?
Perhaps something the WP cron is doing?
Member — 23rd March 2011 (1 year ago) #
Tim, thanks for the response.
We currently don't allow users to sign up for their own blogs either. We have been using WP Super Cache. Like you said, I think the related posts plugin may be the next one I stop and check to see if that helps.
It does sound like it may be a similar issue to yours though as are resources are going crazy at certain times.
Member — 23rd March 2011 (1 year ago) #
No, they haven't been able to pinpoint the issue. They just now said it may be coming from another, smaller WPMU site on our server.
I've removed some plugins that do a lot of accessing of the posts database.
Sadly, they're also taking some time to respond, which is making it hard to really stop and test pieces.
Support Chimp — 23rd March 2011 (1 year ago) #
If this is your server is it a cPanel/WHM server if so you could do your own checking and even terminate high processes as they occur. Those were not fun days. lol
You could look at configuring the CFS to be more strict and if you don't have the CSF firewall then consider getting it installed.
Is your smaller WPMU install more busy with traffic?
Member — 23rd March 2011 (1 year ago) #
We just got a spike of traffic from another site so I asked them if thats the issue and this was their response:
Member — 23rd March 2011 (1 year ago) #
And with that site, my sites crashed again..
Support Chimp — 23rd March 2011 (1 year ago) #
Don't you hate the mysticism. lol
It does indeed sound like the issues we had, spikes which would amass rapidly then occasionally dissipate quickly or get much more intense quickly that you simply could not get into the server even over SSH. Ours was related to traffic, splogs and a bloated install.
Have you you checked the DB and ran any needed repairs, perhaps optimise the tables?
You could just try disabling every plugin for a few days and then turning just one back on every few days to give it time, maybe see if it is one of the plugins.
Member — 23rd March 2011 (1 year ago) #
Yeah, we did repair and optimize the DB.
Might have to try that with the plugins.
Support Chimp — 23rd March 2011 (1 year ago) #
Also did you consider: http://premium.wpmudev.org/project/multi-db
That might help spread the load? I think one of the guys here could comment on that better though.
Support Chimp — 23rd March 2011 (1 year ago) #
Just back to Supercache for moment have you considered:
Donncha explains in this thread if you are unsure what it is: http://wordpress.org/support/topic/plugin-wp-super-cache-lock-down-super-cache
Are all the recommended enabled?
Sales & Support Lead — 30th March 2011 (1 year ago) #
Hiya gbrown88,
Hoping you got this one sorted as it's been awhile since you've posted. I'm marking this as resolved, but if you need further assistance feel free to re-open this thread.
Thanks!
Become a member