Description of problem: Some database operations could be eliminated by keeping a cache for the records. This could shorten the engine response times and drop the database load. For example VDS static records change very rarely, but they are used almost all the time by the engine, the number of such records rarely exceed some hundred, which is still a very small amount of space in the memory.
we should not add caching until we solve jboss clustering aspects. the idea of not having caches in engine (yes, we have a few culprits) is to allow distributing/partitioning the load of the engine to multiple processes.
I know there are some static Map's to cache data. Probably replacing those to any real cache implementations would help to solve the clustering problems. Probably any cache implementation can work on distributed mode.
isn't the infra for this now done with the infinispan patches to close this?
(In reply to comment #3) > isn't the infra for this now done with the infinispan patches to close this? Last time I have seen the cache implementation it was not caching results from DAO's, but I will check again.
(In reply to comment #3) > isn't the infra for this now done with the infinispan patches to close this? Checked again on latest master and cache is not used for cahcing Db records.
http://gerrit.ovirt.org/#/c/20894/ seems related to this BZ Liran please handle
Please see gerrit change http://gerrit.ovirt.org/#/c/20894/
This is an automated message. Re-targeting all non-blocker bugs still open on 3.4.0 to 3.4.1.
*** This bug has been marked as a duplicate of bug 1159621 ***