Solution to ganglia failure to delete images not needed

Environment: ganglia monitoring cluster, where the gmetad process is located, 1 server a, 3 servers where the gmond process is located, B, C, and D;

Objective: To delete images not used in ganglia (using MongoDB as an example)

Symptom: the/etc/ganglia/CONF. d/mongdb. pyconf file corresponding to B, C, and D has been deleted. However, after gmond is restarted, The MongoDB image still exists;

Steps:
1. Stop the gmetad process on a. The image may be stored in the memory. Therefore, you need to stop the gmetad process for a period of time. Please be sure of the time.
/Sbin/service gmetad stop
2. edit/etc/ganglia/gmetad. conf On A, and comment out data_source corresponding to B, C, and D.
3. Go to the/var/lib/ganglia/rrds/directory in A and delete mongdo _ * in the hostname directory corresponding to _ summaryinfo _ and B, C and D respectively _*. RRD image file.
4. Enter B, C, and D respectively to stop the gmond process above. Delete the/etc/ganglia/CONF. d/mongdb. pyconf file.
/Sbin/service gmond stop

Test:
5. Start the gmetad process on.
/Sbin/service gmetad start
6. Modify the/etc/ganglia/gmetad. conf file on a and add data_source corresponding to B, C, and D.
7. Start the gmond process of B, and IE checks whether mongdb images exist.

If the image still exists, restart from 1 to stop gmetad for as long as possible. Otherwise, the gmond process of c and D is started successfully.