hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Matteo Bertozzi (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-8706) Some improvement in snapshot
Date Sat, 08 Jun 2013 06:00:22 GMT

    [ https://issues.apache.org/jira/browse/HBASE-8706?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13678690#comment-13678690
] 

Matteo Bertozzi commented on HBASE-8706:
----------------------------------------

thanks for the patch update, some more notes

ProcedureCoordinator, or all the code under hbase.procedure shouldn't contain any reference
of SnapshotManager or other snapshot stuff... so this constructor should use TIMEOUT_MILLIS_DEFAULT
from this class and not the snapshot one. also you can use this(TIMEOUT_MILLIS_DEFAULT, rpcs,
pool) instead of rewriting the constructor code.
{code}
public ProcedureCoordinator(ProcedureCoordinatorRpcs rpcs, ThreadPoolExecutor pool) {
   this.timeoutMillis = SnapshotManager.SNAPSHOT_TIMEOUT_MILLIS_DEFAULT;
{code}

                
> Some improvement in snapshot
> ----------------------------
>
>                 Key: HBASE-8706
>                 URL: https://issues.apache.org/jira/browse/HBASE-8706
>             Project: HBase
>          Issue Type: Bug
>          Components: snapshots
>    Affects Versions: 0.94.8, 0.95.0
>            Reporter: binlijin
>         Attachments: HBASE-8706-2.patch, HBASE-8706.patch
>
>
> (1)timeout for Procedure can not be configured.
> {code}
> Procedure's timeout
> ProcedureCoordinator
>   final static long TIMEOUT_MILLIS_DEFAULT = 60000;
>    createProcedure(ForeignExceptionDispatcher fed, String procName, byte[] procArgs,
>       List<String> expectedMembers) {
>     // build the procedure
>     return new Procedure(this, fed, WAKE_MILLIS_DEFAULT, TIMEOUT_MILLIS_DEFAULT,
>         procName, procArgs, expectedMembers);
>   }
> RegionServerSnapshotManager:
>   /** Conf key for max time to keep threads in snapshot request pool waiting */
>   public static final String SNAPSHOT_TIMEOUT_MILLIS_KEY = "hbase.snapshot.region.timeout";
>   /** Keep threads alive in request pool for max of 60 seconds */
>   public static final long SNAPSHOT_TIMEOUT_MILLIS_DEFAULT = 60000;
>   public Subprocedure buildSubprocedure(SnapshotDescription snapshot) {
>     long timeoutMillis = conf.getLong(SNAPSHOT_TIMEOUT_MILLIS_KEY,
>         SNAPSHOT_TIMEOUT_MILLIS_DEFAULT);
>     case FLUSH:
>       SnapshotSubprocedurePool taskManager =
>         new SnapshotSubprocedurePool(rss.getServerName().toString(), conf);
>   }
> {code}
> (2)TakeSnapshotHandler
> after snapshotRegions we should call monitor.rethrowException(); to check if there is
exception and if there is we can skip the verifySnapshot
> (3)too much error message when error happened in some place.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message