[
https://issues.apache.org/jira/browse/HBASE-9821?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13802297#comment-13802297
]
Jimmy Xiang commented on HBASE-9821:
------------------------------------
I was thinking to put the RS startcode inside the scanner id. One formula is
(rand int) + (start code last 9 digits). So total it should no more than 19
digits, still a valid long number.
> Scanner id could collide
> ------------------------
>
> Key: HBASE-9821
> URL: https://issues.apache.org/jira/browse/HBASE-9821
> Project: HBase
> Issue Type: Bug
> Reporter: Jimmy Xiang
> Assignee: Jimmy Xiang
> Priority: Minor
>
> We use a random number as our scanner id. In one server, we guarantee that
> the scanner id is unique. However, it is not guaranteed among different
> servers. When a region server restarts quickly, we could run into some
> scanner id collision issue.
> In one of my run:
> {noformat}
> 2013-10-21 22:43:09,071 INFO [RpcServer.handler=2,port=36020]
> regionserver.HRegionServer: Client tried to access missing scanner
> 4305495321392639779
> 2013-10-21 22:43:09,056 INFO [RpcServer.handler=0,port=36020]
> regionserver.HRegionServer: Client tried to access missing scanner
> 4871518173034616791
> 2013-10-21 22:43:09,054 INFO [RpcServer.handler=29,port=36020]
> regionserver.HRegionServer: Client tried to access missing scanner
> 2494346173615963501
> 2013-10-21 22:43:09,046 INFO [RpcServer.handler=28,port=36020]
> regionserver.HRegionServer: Client tried to access missing scanner
> 8522578499834310167
> 2013-10-21 22:43:09,037 INFO [RpcServer.handler=27,port=36020]
> regionserver.HRegionServer: Client tried to access missing scanner
> 6621035169671703961
> 2013-10-21 22:43:09,011 ERROR [RpcServer.handler=20,port=36020]
> regionserver.HRegionServer:
> org.apache.hadoop.hbase.exceptions.OutOfOrderScannerNextException: Expected
> nextCallSeq: 18 But the nextCallSeq got from client: 4470;
> request=scanner_id: 848804760654927372 number_of_rows: 100 close_scanner:
> false next_call_seq: 4470
> at
> org.apache.hadoop.hbase.regionserver.HRegionServer.scan(HRegionServer.java:3030)
> at
> org.apache.hadoop.hbase.protobuf.generated.ClientProtos$ClientService$2.callBlockingMethod(ClientProtos.java:27022)
> at org.apache.hadoop.hbase.ipc.RpcServer.call(RpcServer.java:1979)
> at org.apache.hadoop.hbase.ipc.CallRunner.run(CallRunner.java:90)
> at
> org.apache.hadoop.hbase.ipc.SimpleRpcScheduler.consumerLoop(SimpleRpcScheduler.java:160)
> at
> org.apache.hadoop.hbase.ipc.SimpleRpcScheduler.access$000(SimpleRpcScheduler.java:38)
> at
> org.apache.hadoop.hbase.ipc.SimpleRpcScheduler$1.run(SimpleRpcScheduler.java:110)
> at java.lang.Thread.run(Thread.java:724)
> 2013-10-21 22:43:09,000 INFO [RpcServer.handler=25,port=36020]
> regionserver.HRegionServer: Client tried to access missing scanner
> 4162107982028594792
> {noformat}
> Normally, the nextCallSeq can not be that different between the expected (18)
> and the one sent over from client (4470). It must be because the new server
> happens to use the same scanner id.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
