924060929 commented on code in PR #69255:
URL: https://github.com/apache/doris/pull/69255#discussion_r4374463383
##########
fe/fe-core/src/main/java/org/apache/doris/catalog/CatalogRecycleBin.java:
##########
@@ -342,6 +343,7 @@ private void eraseDatabase(long currentTimeMs, int keepNum)
{
} finally {
writeUnlock();
}
+ FunctionUtil.cleanUdfCache(db);
Review Comment:
Changing the default `auto_suspend` to 14400 seconds mitigates idle pool
retention, but it does not cover the explicitly supported `auto_suspend=-1`
setting. If a BE is temporarily unreachable and the master changes before
cleanup is acknowledged, the pending cleanup exists only in the old FE's
in-memory `AgentTaskQueue`. `replayEraseDatabase()` on the successor removes
the database metadata without reconstructing the cleanup task, so BE recovery
cannot trigger a retry. With `auto_suspend=-1`, the affected per-function
Python pools can remain resident indefinitely. The cleanup intent still needs
to survive master failover until acknowledgement; please cover this with a test
using `auto_suspend=-1`, pending cleanup, master failover, and subsequent BE
recovery.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
