SYSQUERY_POOL_SPILL event in dc_execution engine events: Any concerns?
Good morning to all,
At one of the customer's production cluster, I found several SYSQUERY_POOL_SPILL event. Is this a performance or configuration concern? I searched the forum/and googled for this which was not much help.
4 node cluster with Vertica version 9.3.1 64 GB, 8 core nodes Total RAW data volume 1.5 TB Catalog size ~2GB ~40 schema, 5K tables across all schema, with most having only super projections SYSQUERY pool settings: Memory size: 1GB (Everything else default).
Average/Max memory usage on cluster is 35-40% and once or twice a day peaks at 85-90%.
As a side question how is the SYSDATA pool used?
Any help/pointers will be of great help.