Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
-
Docs Required
Description
After task StartCachesInParallel we can start caches in parallel but GridQueryProcessor is narrow place because it should be start consistently by following reasons:
- checking index to duplicate(and other checking) require one order on every nodes.
- onCacheStart and createSchema contains a lot of mutex.
- maybe it has other reasons.
After this task GridCacheProcessor#prepareStartCaches should be rewrited.
Attachments
Issue Links
- is caused by
-
IGNITE-8006 Starting multiple caches slows down exchange process on joining node
- Resolved