How do we monitor the var/lib/mysql partition use/free space?

[StPaulPete]

I'm looking for a handy way to monitor the utilization of our var/lib/MySQL disk.  I'd like to get an alert if the disk use is above a certain threshold.  ESM has monitoring of the root partition, but not the MySQL disk.  There is an "audit.archiverWarningThreshold" cluster property but setting it only causes a warning to be written in the audit log, and nobody's going to see that.  It also is only watching the database usage and there are other things on the /var/lib/MySQL partition (notably ssgbin_log files) that can cause usage to grow.  I'm trying to find something provided and supported before I resort to a 3rd party monitor such as Nagios.  Searched the articles, etc, and could not find anything.

[Stephen Hughes]

Pete,

 

From previous posts you mentioned that your team likes HTTP output on alerts and such. The ssg ping page located /ssg/ping?systemInfo?node=<node name from Dashboard -> Cluster Status> will actually report back this information along with a bunch of other information around MySQL status, memory usage, MySQL slave status, etc. You could build a policy/service that is triggered either through a crontab or another trigger mechanism (In version 9.0 we have a new scheduler) that parses the information and sends out an alert if something is outside of parameters. The only other method at the moment would be to use a 3rd party monitoring application to pull this information.

 

Sample output from the /ssg/ping page:

********************************************************************************

********************************* Drive Space **********************************

********************************************************************************

Filesystem Size Used Avail Use% Mounted on

/dev/mapper/vg00-lv_root

  7.9G 2.2G 5.4G 29% /

tmpfs 16G 0 16G 0% /dev/shm

/dev/sda1 1008M 55M 902M 6% /boot

/dev/mapper/vg00-lv_home

  2.0G 1.1G 880M 55% /home

/dev/mapper/vg00-lv_opt

  5.0G 2.9G 1.9G 60% /opt

/dev/mapper/vg00-lv_tmp

  2.0G 69M 1.9G 4% /tmp

/dev/mapper/vg00-lv_var

  3.0G 180M 2.7G 7% /var

/dev/mapper/vg00-lv_db

  21G 3.4G 17G 18% /var/lib/mysql

/dev/mapper/vg00-lv_log

  2.0G 120M 1.8G 7% /var/log

/dev/mapper/vg00-lv_audit

  2.0G 95M 1.8G 5% /var/log/audit

 

Sincerely,

 

Stephen Hughes

Director, CA Support

[StPaulPete]

Thanks Steve, I'd already written a script to use the df command and send an http post via curl, but I'll look into this also.  I'm afraid I don't quite know how to interpret your url instruciton.  If I use http://myvip.org.com/ssg/ping?systemInfo?node=MyGateway1, for example, all I get back is "OK".  I'm not sure how to incorporate that cluster status part you mention.

[Anon Anon]

That is most likely because you have are accessing the ping url over http.

According to this: Miscellaneous Cluster Properties - CA API Gateway - 8.3 - CA Technologies Documentation for the pingServlet.mode property:

 

OPEN: Minimal response when request is submitted without SSL (port 8080); full response when request is submitted with SSL (port 8443).

 

If you try to access it over HTTPS you should be able to see the full information Stephen mentions, if you also provide HTTP basic credentials of an administrative user.

Any user with at least the Operator role (read only access to gateway) should do. ( Predefined Roles and Permissions - CA API Gateway - 8.3 - CA Technologies Documentation )

[StPaulPete]

Yes, that worked.  Thanks.

[Legacy User]

if you call the https version of the url, You will get a table with node name at the last column. Click on that and you will be able to see detailed system information.  We use snmp monitoring for receiving alerts on disk space and its very effective.