-
Enhancement
-
Resolution: Done
-
Major
-
6.0.0.Final
-
None
ATM if a tx2 waits on a lock to be released by tx1, the (incoming) tx2 thread is BLOCKED and taken out from the pool. A better way to implement this is by using a structure that would hold and order the depending transactions on the receiver correlated with the async execution API from JGroups 3.3.0, result being that tx2 would not BLOCK but return to the pool this (not maxing out the thread pool that executes commands).
More about it in the "Async invocation section" of https://docspace.corp.redhat.com/docs/DOC-134411
- incorporates
-
ISPN-2986 Intermittent failure to start new nodes during heavy write load
- Resolved
- is blocked by
-
ISPN-5038 Remote package refactor
- Closed
-
ISPN-5370 Make clear() non-transactional and lock free
- Closed
- relates to
-
ISPN-4873 Statetransfer thread pool deadlock
- Closed
-
ISPN-4585 Prioritize commands in the remote executor
- Resolved
-
ISPN-4610 Implement total order for non-transactional caches
- Closed