[
https://issues.apache.org/jira/browse/HIVE-11388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15066851#comment-15066851
]
Alan Gates commented on HIVE-11388:
-----------------------------------
I would suggest adding hostname to the MUTEX_TABLE so that admins can easily
figure out which instance of metastore is running a given thread.
But what happens when that metastore dies? How does one of the existing
metastores know to start an initiator? The table will also need a
'last_hearbeat' column that is updated each time the initiator runs. Other
initiator threads will need to spin checking this every minute or so to see if
the previous initiator has died, and if so take over. If the first metastore
then heartbeats again it would have to discover it no longer holds the lock and
kill its initiator thread.
> there should only be 1 Initiator for compactions per Hive installation
> ----------------------------------------------------------------------
>
> Key: HIVE-11388
> URL: https://issues.apache.org/jira/browse/HIVE-11388
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Affects Versions: 1.0.0
> Reporter: Eugene Koifman
> Assignee: Eugene Koifman
> Priority: Critical
>
> org.apache.hadoop.hive.ql.txn.compactor.Initiator is a thread that runs
> inside the metastore service to manage compactions of ACID tables. There
> should be exactly 1 instance of this thread (even with multiple Thrift
> services).
> This is documented in
> https://cwiki.apache.org/confluence/display/Hive/Hive+Transactions#HiveTransactions-Configuration
> but not enforced.
> Should add enforcement, since more than 1 Initiator could cause concurrent
> attempts to compact the same table/partition - which will not work.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
