Details
-
Bug
-
Status: Stalled (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.4.11, 10.1(EOL), 10.2(EOL), 10.3(EOL), 10.4(EOL)
-
Any
Description
I have project with 8 databases, each 500Mb.
Each has ~800 tables, some of them with triggers.
Use ~100 stored procedures and functions.
Notices excessibe memory usage 128GB is not the limit.
Then HW limit is reached - OOM kills MariaDB
Prepared simple DB structure examples:
dSTAR16_MariaDB.sql (with 16 tables) - uses 2GB
dSTAR64_MariaDB.sql (with 64 tables) - uses 32GB
It is not linear (8 tables would be ~250Mb)
In such DB simple (imposible query) is enougth to massivie memory usage:
UPDATE jj001 SET vj001 = 0 WHERE 'TEJA' = 'FEJA';
|
Could not find server config variables that could limit such memory usage.
Tried script on hosting servers with shared MySQL/MariaDB service - all of them failed using this kind of script.
256 tables script (expected 512GB memory usage) would kill almost any server.
Attachments
Issue Links
- relates to
-
MDEV-19287 Memory leak issue in systemD on mariaDB cluster with remote ssh
- Closed