Grokbase Groups HBase dev June 2008
FAQ
I'm trying to figure out how write ahead logging is working, but don't
see where/how the log is read (eg on regionserver recovery). Can
someone let me know what the status is? Is it intended to wait on hdfs
append before working on recovery process?

cheers,
-clint

Search Discussions

  • Jim Kellerman at Jun 9, 2008 at 7:48 pm
    Logs are recovered by the Master when it detects that a region server has died.
    If you are looking at trunk, see: o.a.h.h.master.BaseScanner.checkAssigned and o.a.h.master.ProcessServerShutdown

    ---
    Jim Kellerman, Senior Engineer; Powerset

    -----Original Message-----
    From: clint.a.m@gmail.com On
    Behalf Of Clint Morgan
    Sent: Monday, June 09, 2008 12:42 PM
    To: hbase-dev@hadoop.apache.org
    Subject: current status of regionserver's logs

    I'm trying to figure out how write ahead logging is working,
    but don't see where/how the log is read (eg on regionserver
    recovery). Can someone let me know what the status is? Is it
    intended to wait on hdfs append before working on recovery process?

    cheers,
    -clint

    No virus found in this incoming message.
    Checked by AVG.
    Version: 8.0.100 / Virus Database: 270.1.0/1492 - Release
    Date: 6/9/2008 10:29 AM
    No virus found in this outgoing message.
    Checked by AVG.
    Version: 8.0.100 / Virus Database: 270.1.0/1492 - Release Date: 6/9/2008 10:29 AM
  • Clint Morgan at Jun 9, 2008 at 8:06 pm
    Cool, so ProcessServerShutdown splits up a reionservers logs amongst
    its regions.

    Where does the region replay the log when it comes back up (to move
    non-flushed edits into Memcache/hstore)?

    On Mon, Jun 9, 2008 at 12:48 PM, Jim Kellerman wrote:
    Logs are recovered by the Master when it detects that a region server has died.
    If you are looking at trunk, see: o.a.h.h.master.BaseScanner.checkAssigned and o.a.h.master.ProcessServerShutdown

    ---
    Jim Kellerman, Senior Engineer; Powerset

    -----Original Message-----
    From: clint.a.m@gmail.com On
    Behalf Of Clint Morgan
    Sent: Monday, June 09, 2008 12:42 PM
    To: hbase-dev@hadoop.apache.org
    Subject: current status of regionserver's logs

    I'm trying to figure out how write ahead logging is working,
    but don't see where/how the log is read (eg on regionserver
    recovery). Can someone let me know what the status is? Is it
    intended to wait on hdfs append before working on recovery process?

    cheers,
    -clint

    No virus found in this incoming message.
    Checked by AVG.
    Version: 8.0.100 / Virus Database: 270.1.0/1492 - Release
    Date: 6/9/2008 10:29 AM
    No virus found in this outgoing message.
    Checked by AVG.
    Version: 8.0.100 / Virus Database: 270.1.0/1492 - Release Date: 6/9/2008 10:29 AM
  • Clint Morgan at Jun 9, 2008 at 8:10 pm
    Nevermind, found it in HStore
    On Mon, Jun 9, 2008 at 1:05 PM, Clint Morgan wrote:
    Cool, so ProcessServerShutdown splits up a reionservers logs amongst
    its regions.

    Where does the region replay the log when it comes back up (to move
    non-flushed edits into Memcache/hstore)?

    On Mon, Jun 9, 2008 at 12:48 PM, Jim Kellerman wrote:
    Logs are recovered by the Master when it detects that a region server has died.
    If you are looking at trunk, see: o.a.h.h.master.BaseScanner.checkAssigned and o.a.h.master.ProcessServerShutdown

    ---
    Jim Kellerman, Senior Engineer; Powerset

    -----Original Message-----
    From: clint.a.m@gmail.com On
    Behalf Of Clint Morgan
    Sent: Monday, June 09, 2008 12:42 PM
    To: hbase-dev@hadoop.apache.org
    Subject: current status of regionserver's logs

    I'm trying to figure out how write ahead logging is working,
    but don't see where/how the log is read (eg on regionserver
    recovery). Can someone let me know what the status is? Is it
    intended to wait on hdfs append before working on recovery process?

    cheers,
    -clint

    No virus found in this incoming message.
    Checked by AVG.
    Version: 8.0.100 / Virus Database: 270.1.0/1492 - Release
    Date: 6/9/2008 10:29 AM
    No virus found in this outgoing message.
    Checked by AVG.
    Version: 8.0.100 / Virus Database: 270.1.0/1492 - Release Date: 6/9/2008 10:29 AM

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupdev @
categorieshbase, hadoop
postedJun 9, '08 at 7:43p
activeJun 9, '08 at 8:10p
posts4
users3
websitehbase.apache.org

People

Translate

site design / logo © 2022 Grokbase