[
https://issues.jboss.org/browse/ISPN-2366?page=com.atlassian.jira.plugin....
]
Adrian Nistor commented on ISPN-2366:
-------------------------------------
I have a case in a functional test for ISPN-2362 that is very hard to create unless
CacheManager.getCache() returns ASAP instead of waiting for NBST to perform the transfer.
I'm not saying we should make getCache() always return immediately but we could make
this configurable, some users might need it, most will probably prefer the current
behaviour. So this whole issue can be solved by introducing a boolean config
'waitForInitialStateTransferToComplete' in StateTransferConfiguration (and
friends) that is 'true' by default and updating StateTransferManagerImpl to skip
the waiting if configuration says so. I already implemented this in my t_2362_m branch if
you are interested. Would you agree to include this in 5.2.0?
Allow getCache() to return before the node receives any state
-------------------------------------------------------------
Key: ISPN-2366
URL:
https://issues.jboss.org/browse/ISPN-2366
Project: Infinispan
Issue Type: Task
Components: State transfer
Affects Versions: 5.2.0.Alpha4
Reporter: Dan Berindei
Assignee: Adrian Nistor
Fix For: 6.0.0.Final
With NBST it's possible to use a cache without having any local state. We should
allow getCache() to return as soon as the joiner has received a consistent hash (actually
CacheTopology) from the coordinator.
This is related to ISPN-1394 - joiners should be able to work even if rebalancing is
blocked.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:
http://www.atlassian.com/software/jira