J-D and Andy,
This seems to solve the problem. I thought I had set this parameter
before but realized I set the "master" lease time instead of the
"region server" lease time.
The MR task makes http calls, so I also needed to set the timeout on
the call to make sure it doesn't take longer than the ping back to the
server.
get = new GetMethod(url);
get.getParams().setParameter(HttpMethodParams.RETRY_HANDLER, new
DefaultHttpMethodRetryHandler(1, false));
get.getParams().setParameter(HttpMethodParams.SO_TIMEOUT, 10000);
get.getParams().setParameter("http.connection.stalecheck", true);
get.getParams().setParameter("http.connection.timeout", 30000);
Finally, the MR task loops through the result set and stores them in a
hbase table. I added the report.progress() call to let the server
know its still working.
Thanks for your help,
Dru
On Aug 11, 2008, at 4:57 PM, Andrew Purtell wrote:
I use 120 seconds.
The patch for HBASE-816 may work for you also:
https://issues.apache.org/jira/secure/attachment/12387963/816.v2.patch
- Andy
From: Jean-Daniel Cryans <[EMAIL PROTECTED]>
Subject: Re: Unknown Scanner Exception
To: [email protected]
Date: Monday, August 11, 2008, 4:06 PM
Dru,
It may be also caused by a compaction going on (or a flush).
Best way to be sure is to up the regionserver lease timeout.
Default is 30 seconds but I recommend 60 seconds for MR
jobs. See if this works for you.
J-D
On Mon, Aug 11, 2008 at 6:34 PM, Dru Jensen
<[EMAIL PROTECTED]> wrote:
Hi J-D,
I watched as one of the map tasks completed
successfully. Another one was
launched as a child of the second MR task. then I get
the Unknown Scanner
Exception:
org.apache.hadoop.hbase.UnknownScannerException:
org.apache.hadoop.hbase.UnknownScannerException: Name:
-463563612898654250
[...]
