[
https://issues.apache.org/jira/browse/HBASE-11165?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14098801#comment-14098801
]
Mikhail Antonov commented on HBASE-11165:
-----------------------------------------
(not sure it's right to discuss it here, may be better elsewhere?) but that
urged me to look at those methods, and I'm wondering, if the invariants there
allow more fine-grained locking (sorry, may be it was already asked elsewhere
and I missed conversation). Like:
- RegionStates#getRegionState(String) is synchronized, and does lookup in
HashMap. Should it be ConcurrentHashMap instead?
- in RegionStates#updateRegionState, regionsInTransition and regionsStates do
form invariant, what about lastAssignments? Can updating lastAssignments happen
under separate subsequent lock?
[~jxiang] what do you think?
> Scaling so cluster can host 1M regions and beyond (50M regions?)
> ----------------------------------------------------------------
>
> Key: HBASE-11165
> URL: https://issues.apache.org/jira/browse/HBASE-11165
> Project: HBase
> Issue Type: Brainstorming
> Reporter: stack
> Attachments: HBASE-11165.zip, Region Scalability test.pdf,
> zk_less_assignment_comparison.pdf
>
>
> This discussion issue comes out of "Co-locate Meta And Master HBASE-10569"
> and comments on the doc posted there.
> A user -- our Francis Liu -- needs to be able to scale a cluster to do 1M
> regions maybe even 50M later. This issue is about discussing how we will do
> that (or if not 50M on a cluster, how otherwise we can attain same end).
> More detail to follow.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
