Disk Space Optimization
To prevent running out of disk space, you can optimize disk space utilization. I.e., you can accumulate less data, or keep the data for a shorter period, without missing what is important for your cluster operation. Data collection and accumulation parameters should be defined in the optimal way at the beginning - right after the Control Center installation.
If you do run out of disk space in the course of Control Center operation, this might mean that your data collection and accumulation parameters need further optimization. Proceed as follows.
Limit Saved Data
Combine the following approaches:
Throttle Down Data Collection
Clean Up Space (Optional)
If you need to not only prevent disk clogging in the future but free up space that is already occupied, proceed as follows:
Identify Oversized Tables
Identify the tables that occupy the most space on your disk:
-
Check the sizes of the relevant tables on the disk - QuerySession, Trace, Span, and TaskSession - in
work/db/ggcc_db. Because of defragmentation, the table might appear significantly larger than in reality. For defragmented tables, check the actuator metrics for real sizes. -
Get the number of records in the same tables by from Spring actuator metrics at the following URLs:
-
{cc_base_url}/actuator/metrics/repository.table.Span.size
-
{cc_base_url}/actuator/metrics/repository.table.Trace.size
-
{cc_base_url}/actuator/metrics/repository.table.QuerySession.size
-
{cc_base_url}/actuator/metrics/repository.table.TaskSession.size
-
If the relevant tables in your environment are not yet of representative size, you can use our estimates for a 3-node cluster with an average load:
Delete Oversized Tables
For optimization settings to take effect, you need to clean up the oversized tables. Because of defragmentation, each table occupies the maximum disk space it had achieved in its lifecycle. The occupied size will remain the same, even if the table becomes "empty," unless you physically delete the defragmented table from the disc.
To clean up the tables:
-
Stop Control Center.
-
Delete the oversized table (cache) folder. For example, delete
work/db/ggcc_db/cache-QuerySessionCacheto clean up the space occupied by the running query data. -
Start Control Center.
© 2024 GridGain Systems, Inc. All Rights Reserved. Privacy Policy | Legal Notices. GridGain® is a registered trademark of GridGain Systems, Inc.
Apache, Apache Ignite, the Apache feather and the Apache Ignite logo are either registered trademarks or trademarks of The Apache Software Foundation.