[jira] Commented: (HBASE-669) MultiRegion transactions with Optimistic Concurrency Control

["stack (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-669?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12603966#action_12603966 ] 

stack commented on HBASE-669:
-----------------------------

Couple of quick comments in no particular order after cursory read (so some of the questions might be extra dumb):
 
+ The client runs the delete edits to invalidate the transaction?  (By replaying the edits out of the client-side WAL)?  If so, can it be done regionserver side?  If client lease expires or client cancels the transaction, the regionserver does the undo operations?
+ Good stuff making HTable, etc. subclassable.  If you want to do these changes as a separate patch, I'll just commit it so you don't have to reproduce it in this and your secondary index patch.  Make an issue.
+ On, feedback regards how to have more than one regionserver implementation -- transactional and secondary indices -- I think yeah, probably has to be some mixin factility.  I can see this transactional stuff eventually making it into core.
+ I don't get the global transaction log doohickey.  Why not just have regionservers just discard COMMIT_PENDING if lease times out or on restart, comes across a COMMIT_PENDING replaying?
+ We're trying to get zookeeper into the mix tout-de-suite.  Then you wouldn't have to do a custom server?
+ Classes are missing licenses.  Some of those that have them, have 2007 as date.
+ On '// FIXME, not sure why this is needed, but copied over from HTable', you want to use Arrays.toArray or some such? (I ain't sure why its there either)
+ Why pass transaction state when doing scan/get reads?
+ Add a package info under client transactional that explains what this feature is, how it works and how to use it (will help others get invovled).
+ TransctionalState on server side should inherit from the client TransactionState?  They have mostly same enums?
+ I applied your getBytes.toInt ...HBASE-676 (thanks)







> MultiRegion transactions with Optimistic Concurrency Control
> ------------------------------------------------------------
>
>                 Key: HBASE-669
>                 URL: https://issues.apache.org/jira/browse/HBASE-669
>             Project: Hadoop HBase
>          Issue Type: New Feature
>          Components: client, ipc, regionserver
>            Reporter: Clint Morgan
>         Attachments: hbase-669-v2.patch, hbase-669.patch
>
>
> We have a need for ACID transactions across tables. This issue is about adding transactions which span multiple regions.  We do not envision many competing writes, and will be read-dominated in general. This makes Optimistic Concurrency Control (OCC) seem like the way to go.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.