Bit of a newb when it comes to sysadmin stuff - would greatly appreciate some help:

Details:

Running Debian 7 Latest 64 bit on a 24320 MB Linode

Running apache2 to serve up a php 5.4.4 web app -

Installed Piwik via the Linode documentation - installed mysql in the course of doing so - everything ran smoothly for about 3 months - never any problems or slowness.

Added a couple of long running node scripts on the same box.

Got a new relic alert that memory usage was over 95% - assumed this was wrong, linode manager shows an average of 4% cpu - nothing on the box is really running at a scale yet that should cause any problems.

Then today I go to hit my piwik analytics page and I get an error: cannot connect to mysql - so mysql crashed - was likely responsible for the thrashing now visible on linode manager io panel.

Go to try and start mysql manually= [FAIL] init.d/mysql: ERROR: The partition with /var/lib/mysql is too full! ... failed!

Filesystem Size Used Avail Use% Mounted on
rootfs 24G 6.3G 16G 29% /
/dev/root 24G 6.3G 16G 29% /
devtmpfs 493M 0 493M 0% /dev
tmpfs 99M 184K 99M 1% /run
tmpfs 5.0M 0 5.0M 0% /run/lock
tmpfs 249M 0 249M 0% /run/shm

Filesystem Inodes IUsed IFree IUse% Mounted on
rootfs 1319360 83692 1235668 7% /
/dev/root 1319360 83692 1235668 7% /
devtmpfs 126139 1442 124697 2% /dev
tmpfs 126406 793 125613 1% /run
tmpfs 126406 3 126403 1% /run/lock
tmpfs 126406 2 126404 1% /run/shm

Meanwhile there is nothing in my var/log/mysql - all the mysql related log files are empty

I really liked Piwik, but from browsing around it seems other Linode users have had issues with it - and there's talk of a memory leak.

I would love to have analytics that are not harvesting all my data for the empire (google) but even more importantly I'd like to know what happened here so I can prevent it in the future.

Any thoughts or suggestions?

*cricket*

nobody?

Are you sure piwik is the culprit that's causing your memory issue (which is different from your cpu usage of 4%) or your disk space issue? We've been using piwik for a few years now and haven't had any issues with it.

If you think it's piwik have you looked at your table sizes & record counts? Are they high? Have you archived some of your activity (I forget what they call the processes; our designer takes care of the piwik admin)?

A few more details - now I am not so sure.

After running Piwik for a while I added a wordpress install - which is also hosed now due to not being able to start mysql

Not sure what to do about being unable to start mysql and getting a space error despite having plenty of disk space...

Mmm, sudo apt-get clean and then sudo /etc/init.d/mysql restart seems to have done the trick - need to figure out what originally happened, could wordpress and piwik be colliding bc of a setting I have misconfigured?

Wordpress and piwik should be fine together.

MySQL can (and will) die if it runs out of RAM. What does your MySQL log say (from around the time of the problems)? How big of a Linode are you running? What do you get when you run "free -m"?

Thanks for your help:

total used free shared buffers cached
Mem: 987 971 16 0 1 26
-/+ buffers/cache: 943 44
Swap: 255 255 0

Still havent claimed my full sized linode yet.

All of my mysql logs are empty, including the one explicitly set in my.cnf as well as syslog

With those numbers it looks like you're running out of memory & swap space. You should look at the same things that pex23 needs to in this thread from earlier today, such as Apache & kworker processes.

Very much appreciated.

zack2121,
I forgot about this potential RAM issue with Wordpress. If you're running Wordpress it could be the 'Max RAM' setting isn't defined so it tries to take 256M. Look at this thread for more info.