[
https://issues.apache.org/jira/browse/HBASE-15143?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15923066#comment-15923066
]
stack commented on HBASE-15143:
-------------------------------
It looks good [~balazs.meszaros] Is the test failure yours? Thanks for the
screen shots.
LockInfo[] listLocks() matches listProcedures..... Good.
We have to do this:
223 <dependency>
224 <groupId>org.apache.hbase</groupId>
225 <artifactId>hbase-protocol-shaded</artifactId>
226 </dependency>
Hmm.... Was trying to have it so hbase-common did NOT depend on
hbase-protocol* but I see the procedure stuff already does this. We can file a
follow-on JIRA to purge protos from hbase-common. Not your fault. But let me
ask, you are modeling on the ProcedureInfo? if so, that is a bit of a hack.
Dang. Would be good to do a right-soln for ProcedureInfo.... and therefore for
LockInfo. But this is not your issue.
Should WaitingProcedure be a ProcedureInfo altogether? OR a subclass? Or have-a
PI? Could you have LockInfo protobuf use PI protobuf?
LockInfo needs a license. See other source files. And TestLockUtil needs
license. Maybe should be in hbase.procedure2 package?
Ditto for LockUtil
LockUtil and LockInfo are in the top-level of the hbase package. Is that right?
I suppose PI is and you are following that precedent.
Otherwise, patch looks great.
> Procedure v2 - Web UI displaying queues
> ---------------------------------------
>
> Key: HBASE-15143
> URL: https://issues.apache.org/jira/browse/HBASE-15143
> Project: HBase
> Issue Type: Sub-task
> Components: proc-v2, UI
> Reporter: Matteo Bertozzi
> Assignee: Balazs Meszaros
> Priority: Minor
> Fix For: 2.0.0
>
> Attachments: HBASE-15143-BM-0001.patch, HBASE-15143-BM-0002.patch,
> HBASE-15143-BM-0003.patch, HBASE-15143-BM-0004.patch, screenshot.png
>
>
> We can query MasterProcedureScheduler to display the various procedures and
> who is holding table/region locks.
> Each procedure is in a TableQueue or ServerQueue, so it is easy to display
> the procedures in its own group.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
