Details
-
Bug
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Won't Fix
-
1.5.2
-
None
-
centos 7
32G RAM, 6 cores
Description
setup is a single node mariadb 10.5.4 with columnstore 1.5.2
during and ETL job sequentially chaining insert-select bulk load statements, we can see memory usage increase until around 90%
Then statements fail with the error : Error in forking cpimport.bin (errno-12); Cannot allocate memory
job is written in java, using a connection pool.
jemalloc fix has been applied already
Innodb buffer pool is 25% of total server memory.
TotalUmMemory + NumBlocksPct is 75%
Attachments
Issue Links
- duplicates
-
MCOL-3508 Error in forking cpimport.bin (errno-12); Cannot allocate memory
- Closed