[
https://issues.apache.org/jira/browse/HBASE-10926?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13987246#comment-13987246
]
Jerry He commented on HBASE-10926:
----------------------------------
bq. you don't reject concurrent flush on the same table, is that on purpose?
Concurrent flush on the same table is rejected.
{code}
// We rely on the existing Distributed Procedure framework to prevent any
concurrent
// procedure with the same name.
Procedure proc = coordinator.startProcedure(monitor, desc.getInstance(),
new byte[0], Lists.newArrayList(regionServers));
monitor.rethrowException();
if (proc == null) {
String msg = "Failed to submit distributed procedure " +
desc.getSignature() + " for '"
+ desc.getInstance() + "'. " + "Another flush procedure is running?";
LOG.error(msg);
throw new IOException(msg);
}
{code}
In the master log:
{code}
2014-05-01 11:15:26,630 INFO org.apache.hadoop.hbase.master.HMaster:
Client=biadmin//9.30.194.18 procedure request for: flush-table-proc
2014-05-01 11:15:26,644 WARN
org.apache.hadoop.hbase.procedure.ProcedureCoordinator: Procedure table2
currently running. Rejecting new request
2014-05-01 11:15:26,644 ERROR
org.apache.hadoop.hbase.procedure.ProcedureCoordinator: Failed to submit
procedure 'table2'
2014-05-01 11:15:26,644 ERROR
org.apache.hadoop.hbase.procedure.flush.MasterFlushTableProcedureManager:
Failed to submit distributed procedure flush-table-proc for 'table2'. Another
flush procedure is running?
2014-05-01 11:15:26,852 INFO org.apache.hadoop.hbase.master.HMaster:
Client=biadmin//9.30.194.18 procedure request for: flush-table-proc
2014-05-01 11:15:26,869 WARN
org.apache.hadoop.hbase.procedure.ProcedureCoordinator: Procedure table2
currently running. Rejecting new request
2014-05-01 11:15:26,869 ERROR
org.apache.hadoop.hbase.procedure.ProcedureCoordinator: Failed to submit
procedure 'table2'
2014-05-01 11:15:26,869 ERROR
org.apache.hadoop.hbase.procedure.flush.MasterFlushTableProcedureManager:
Failed to submit distributed procedure flush-table-proc for 'table2'. Another
flush procedure is running?
{code}
Interestingly, in HBaseAdmin executeCallable, which execProcedure() uses , is
RPC with retries. So in the log you can see the client just keeps trying until
the previous flush procedure completes.
Then it will be able to run on the same table.
{code}
2014-05-01 11:15:38,788 INFO org.apache.hadoop.hbase.master.HMaster:
Client=biadmin//9.30.194.18 procedure request for: flush-table-proc
2014-05-01 11:15:38,804 DEBUG
org.apache.hadoop.hbase.procedure.ProcedureCoordinator: Procedure table2 was in
running list but was completed. Accepting new attempt.
2014-05-01 11:15:38,804 INFO org.apache.hadoop.hbase.procedure.Procedure:
Starting procedure 'table2'
2014-05-01 11:15:38,804 DEBUG
org.apache.hadoop.hbase.errorhandling.TimeoutExceptionInjector: Scheduling
process timer to run in: 60000 ms
2014-05-01 11:15:38,804 DEBUG org.apache.hadoop.hbase.procedure.Procedure:
Procedure 'table2' starting 'acquire'
...
{code}
So effectively concurrent requests from the client will be executed
sequentially.
> Use global procedure to flush table memstore cache
> --------------------------------------------------
>
> Key: HBASE-10926
> URL: https://issues.apache.org/jira/browse/HBASE-10926
> Project: HBase
> Issue Type: Improvement
> Components: Admin
> Affects Versions: 0.96.2, 0.98.1
> Reporter: Jerry He
> Assignee: Jerry He
> Fix For: 0.99.0
>
> Attachments: HBASE-10926-trunk-v1.patch, HBASE-10926-trunk-v2.patch
>
>
> Currently, user can trigger table flush through hbase shell or HBaseAdmin
> API. To flush the table cache, each region server hosting the regions is
> contacted and flushed sequentially, which is less efficient.
> In HBase snapshot global procedure is used to coordinate and flush the
> regions in a distributed way.
> Let's provide a distributed table flush for general use.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
