OpsCenter 3.2 introduced an issue that caused metrics data to never expire. This old data will not go away on its own. In order to free the space, you will want to follow the workaround instructions.
The data in the OpsCenter Keyspace will continue to grow and take up increasingly more disk space over time.
This is due to a bug in our clojure code which caused metrics data to be inserted without a ttl.
If your OpsCenter keyspace is using too much space, you have two options for a workaround.
- If you are fine with discarding historical metrics data, you can perform a truncate on any of the column families that have grown too large. The most common offenders will be the pdps (raw data points) and rollups60 (1m data points) column families. To truncate these column families, using cqlsh 3:
- use "OpsCenter";
- TRUNCATE rollups60;
- TRUNCATE pdps;
- If you would like to keep some of the historical metrics data you will want to manually remove any sstables who's last modified time is older than the target expire datetime. Once you have removed the sstables we recommend that you restart dse/c*.
This issue is resolved in OpsCenter 3.2.1 with an expected release the week of August 11, 2013. Please note, however, that upgrading to to 3.2.1 will resolve the issue going forward but is not retroactive. You will still want to take action with one of the two workarounds for the data collected while running OpsCenter 3.2.
OpsCenter Release Notes