hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kannan Muthukkaruppan (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-2365) Double-assignment around split
Date Thu, 25 Mar 2010 20:47:27 GMT

    [ https://issues.apache.org/jira/browse/HBASE-2365?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12849895#action_12849895
] 

Kannan Muthukkaruppan commented on HBASE-2365:
----------------------------------------------

BaseScanner.java:checkAssigned() has some comments related to double-assignment:

{code}
   // Scans are sloppy. They don't respect row locks and they get and
    // cache a row internally so may have data that is stale. Make sure that for
    // sure we have the right server and servercode. We are trying to avoid
    // double-assignments. See hbase-1784. Will have to wait till 0.21 hbase
    // where we use zk to mediate state transitions to do better.
{code}

And it seems to deal with the staleness of the scanner by doing a GET for the relevant key.
But the following line:

{code}
       // Now get the region assigned
        this.master.regionManager.setUnassigned(info, true);
{code}

uses the potentially stale "info" passed as an argument to checkAssigned(). Should this instead

retrieve the HRegionInfo from the result of the GET?


> Double-assignment around split
> ------------------------------
>
>                 Key: HBASE-2365
>                 URL: https://issues.apache.org/jira/browse/HBASE-2365
>             Project: Hadoop HBase
>          Issue Type: Bug
>            Reporter: stack
>            Priority: Blocker
>             Fix For: 0.20.4, 0.21.0
>
>         Attachments: zheng.tgz
>
>
> Its looking like we have a split updating .META. with daughter regions and then before
we process the split in master, one of the daughters has already been assigned.  On processing
of the split, we assign daughter again.
> I thought this had been fixed previously?  Doesn't seem so.  Need to look again.
> Here is evidence for region named:
> {code}summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017{code}
> First master-side:
> {code}
> 2010-03-15 16:06:52,153 INFO org.apache.hadoop.hbase.master.RegionManager: Assigning
region summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
to cactus208,60020,12686305486412010-03-15 16:06:52,156 INFO org.apache.hadoop.hbase.master.BaseScanner:
RegionManager.metaScanner scan of 245 row(s) of meta region {server: 172.16.1.209:60020, regionname:
.META.,,1, startKey: <>} complete2010-03-15 16:06:52,156 INFO org.apache.hadoop.hbase.master.BaseScanner:
All 1 .META. region(s) scanned2010-03-15 16:06:52,841 INFO org.apache.hadoop.hbase.master.RegionManager:
Assigning region summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E5\x25A4\x25A7\x25E8\x25BF\x259E\x25E5\x2588\x2598\x25E5\x25B8\x2588\x25E5\x2582\x2585\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E5\x2587\x25B9\x25E9\x2599\x25B7\x25E4\x25BF\x25AE\x25E5\x25A4\x258D\x25E6\x2596\x25BD\x25E5\x25B7\x25A5\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E7\x2594\x25A8\x25E5\x2593\x2581\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
to cactus210,60020,12686305508862010-03-15 16:06:54,377 INFO org.apache.hadoop.hbase.master.ServerManager:
Processing MSG_REPORT_SPLIT: summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E5\x25A4\x25A7\x25E8\x25BF\x259E\x25E5\x2588\x2598\x25E5\x25B8\x2588\x25E5\x2582\x2585\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E5\x2587\x25B9\x25E9\x2599\x25B7\x25E4\x25BF\x25AE\x25E5\x25A4\x258D\x25E6\x2596\x25BD\x25E5\x25B7\x25A5\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E7\x2594\x25A8\x25E5\x2593\x2581\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268611349836:
Daughters; summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E5\x25A4\x25A7\x25E8\x25BF\x259E\x25E5\x2588\x2598\x25E5\x25B8\x2588\x25E5\x2582\x2585\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E5\x2587\x25B9\x25E9\x2599\x25B7\x25E4\x25BF\x25AE\x25E5\x25A4\x258D\x25E6\x2596\x25BD\x25E5\x25B7\x25A5\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E7\x2594\x25A8\x25E5\x2593\x2581\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017,
summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
from cactus209,60020,1268630548451; 1 of 32010-03-15 16:06:54,388 INFO org.apache.hadoop.hbase.master.RegionManager:
Assigning region summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
to cactus209,60020,1268630548451
> {code}
> Its hard to read but above is an assignment, the split message, then what seems to be
same region being assigned again.
> Here is RS side on 209 server:
> {code}
> 2010-03-15 16:06:29,727 INFO org.apache.hadoop.hbase.regionserver.HRegionServer: Worker:
MSG_REGION_OPEN: summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
> 2010-03-15 16:06:29,792 INFO org.apache.hadoop.hbase.regionserver.HRegion: region summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017/1011052036
available; sequence id is 199443346
> 2010-03-15 16:06:29,792 INFO org.apache.hadoop.hbase.regionserver.HRegionServer: Worker:
MSG_REGION_OPEN: summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E5\x25A4\x25A7\x25E8\x25BF\x259E\x25E5\x2588\x2598\x25E5\x25B8\x2588\x25E5\x2582\x2585\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E5\x2587\x25B9\x25E9\x2599\x25B7\x25E4\x25BF\x25AE\x25E5\x25A4\x258D\x25E6\x2596\x25BD\x25E5\x25B7\x25A5\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E7\x2594\x25A8\x25E5\x2593\x2581\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
> 2010-03-15 16:06:29,793 INFO org.apache.hadoop.hbase.regionserver.HRegion: Starting compaction
on region summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017
> 2010-03-15 16:06:29,944 INFO org.apache.hadoop.hbase.regionserver.HRegion: region summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E5\x25A4\x25A7\x25E8\x25BF\x259E\x25E5\x2588\x2598\x25E5\x25B8\x2588\x25E5\x2582\x2585\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E5\x2587\x25B9\x25E9\x2599\x25B7\x25E4\x25BF\x25AE\x25E5\x25A4\x258D\x25E6\x2596\x25BD\x25E5\x25B7\x25A5\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E7\x2594\x25A8\x25E5\x2593\x2581\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017/1971466363
available; sequence id is 199443345
> 2010-03-15 16:06:32,750 INFO org.apache.hadoop.hbase.regionserver.HRegionServer: MSG_REGION_CLOSE_WITHOUT_REPORT:
summary,SITE_0000000032\x01pt\x0120100314000000\x01\x25E7\x258C\x25AE\x25E5\x258E\x25BF\x25E5\x2586\x2580\x25E9\x25B9\x25B0\x25E6\x2591\x25A9\x25E6\x2593\x25A6\x25E6\x259D\x2590\x25E6\x2596\x2599\x25E5\x258E\x2582\x2B\x25E6\x25B1\x25BD\x25E8\x25BD\x25A6\x25E9\x2585\x258D\x25E4\x25BB\x25B6\x25EF\x25BC\x258C\x25E5\x2598\x2580\x25E9\x2593\x2583\x25E9\x2593\x2583--\x25E7\x259C\x259F\x25E5\x25AE\x259E\x25E5\x25AE\x2589\x25E5\x2585\x25A8\x25E7\x259A\x2584\x25E7\x2594\x25B5\x25E8\x25AF\x259D\x25E3\x2580\x2581\x25E7\x25BD\x2591\x25E7\x25BB\x259C\x25E4\x25BA\x2592\x25E5\x258A\x25A8\x25E4\x25BA\x25A4\x25E5\x258F\x258B\x25E7\x25A4\x25BE\x25E5\x258C\x25BA\x25EF\x25BC\x2581,1268640385017:
Duplicate assignment
> {code}
> See how we end with 'Duplicate assignment' message?

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


Mime
View raw message