WordPress MU

Hi,

We have a blogging site online.

We are running WPMU 2.9.1.1 on an Apache VPS with 8gb RAM and 2 x Intel Xeon CPU E5540 2.53Ghz.
Mysql DB and Apache server running on the same machine.
Mysql DB is not that large, approx 150 mb.
We have about 25 000 - 30 000 visits/day

We have at most about 5-6 bloggers online and blogging.
Admin runs in normal speed, system is only 2-5% loaded and have most of the time around 2-4 gb free RAM memory (of the total 8gb RAM memory).
Disk and bandwidth is no problem.
The site for visitors are a bit slow normally for visitors, www-address can be sent if you like to help out.

When checking the server via SSH we can see normally 1 active process running, the Apache process(www-data) when I run the command "top -ci".

Problem accouter when one blogger goes in to publish/change a post, when they are done and click publish or update all the suddenly 10-20 new Apache process(www-data) starts, the mysql write process starts and goes mad consuming all the power of the server.
Mysql process takes 140-180%(we have soft limit of 100% and hard limit of 200%) of the power from the system and the server are around 90-99.9% loaded in total.
Still there is 0.5-2gb free of the RAM memory.

When this happens the publish of the new/changed post takes long time, perhaps 1-5 minutes before the post is live/changed and the sites of course become very slow, the mysql process and the many Apache processes keeps running for approx 10-15 minutes before it is done and everything is back to normal.

Same thing happens every single time a blogger is in and change/adds a post.

Please help!!
What can we do to improve our system?
Is this a normal being of WPMU that the publishing of posts take this much power?

All help is apprichiated!

Thanks in advance,
R

That isn't normal. What happens if you disable all plugins (WPMU and standard WP ones)?

Might a health check on the database tables be worthwhile, to rule out any damaged tables? Not sure how that might cause symptoms, but wonder if you can get database retries going on, consuming resources...

Hi,

Thank you for the inputs.
We are right now converting the table of the db to innodb in hope that will speed up things a little bit.
We have also run "repair" on the tables in a try to optimize them.

We also at a separate test tried to disable all the plugins but there where not much change sadly :-(

We have also gone through all the settings on both the Apache server and the mysql server and optimized them as good as we can...

Here are some of the settings we change on mysql, anything that might look odd/low for our size of site?
key_buffer = 64M
join_buffer_size = 2M
max_allowed_packet = 16M
thread_stack = 128K
thread_cache_size = 8
max_connections = 250
tmp_table_size = 256M
max_heap_table_size = 256M
table_cache = 1024

Any other ideas? I am clueless...

Might it be time to upgrade the hardware and have one dedicated server for the DB and one for Apache.
I know this is always a possibility, but not very cost-friendly so I rather check other options first.

Thanks in advance,
R

have you run any of the mysql tuning scripts over the DB?

Also how is Apache configured?

Plugins also mean anything you have in the mu-plugins folder. List them if you have to.

No, this is not normal. 99% of the time issues like this are plugins.