When getting an alert that the System Disk is Full, or potentially seeing some MYSQL errors from collectord, it can mean that the drive or some partition is full and needs to be increased.
Disk space is important in the Exinda due to the back end needing storage for all the monitoring data. All types of monitoring and reporting information, as well as logging and processing information is stored in the Exinda and as a result, depending on the appliance, the various partitions indicating the different storage types can get full, and a message that the "System Disk is Full", can be generated. This usually leads to some monitoring graphs becoming completely unavailable.
There are two possible causes for this issue:
1.- Ever since version 6.4.3, Exinda improved its mechanisms to gather reporting data. This leads to the monitor partition to be full as the 10GB that are given to it by default is no longer enough. You can verify this setting under System-->Setup-->Storage. In this case we should increase this partition to 100GB
2.- If 100GB is not enough and the partition is still filling up, check for the "URL logging" setting under System-->Setup-->Monitoring. This is a setting that is enabled if the SQL database is being accessed by a remote SQL repository (via OleDB or something similar). The URL graphs within the Exinda WebUI do not change with this setting, so the recommendation would be to disable this setting or to reduce the amount of days for which this data is storage. At this point, it might be necessary to clear the database to bring everything back to normal after disabling this feature, for this a support case might be needed.
Please note: There was a bug in URL Logging in the ASAM (System -> Setup -> Monitoring) where users would specify the number of days for which URL logs should be stored, but Exinda was not deleting the old data. It simply kept on adding to the old data hence causing massive increase in database size.
This was a bug up until version 6.4.3U7. It was fixed in version 6.4.3U8 and above.
For the first case, it is possible to resize the partitions through the web UI.
- Go to System-->Setup-->Storage,
- Find an underused partition (if running an x700 license that does not contain acceleration, the WAN Memory partition is perfect, or if edge cache is not enabled, using the edge-cache partition would work).
- Change the size of the partition to a value that is greater than the 'Used' value, but reduces the size by a good amount, and click on "Resize".
- After the status of the partition changes back to "Avaiable" you should see however much space was removed from the partition show up in 'unallocated storage'.
- Change the size of the monitor partition with 90GB more, click on "Resize" and wait again.
It is recommend on waiting 5 minutes for the graphs to be back, if they are not back, we recommend a reboot of the Exinda appliance.
2.- For the second case, disable "URL Logging" under System-->Setup-->Monitoring and open a support case so the database can be cleared.
1) If the var partition of the box is full then Please use the following KB article:/var partition still 100% , unable to find files taking up the space
2) If the issue is related to the /var/data/monitor Partition: Please follow the System Disk Full /var/data/monitor