hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Suresh Srinivas (Issue Comment Edited) (JIRA)" <j...@apache.org>
Subject [jira] [Issue Comment Edited] (HDFS-2301) Start/stop appropriate namenode internal services during transition to active and standby
Date Wed, 28 Sep 2011 23:25:45 GMT

    [ https://issues.apache.org/jira/browse/HDFS-2301?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13116699#comment-13116699
] 

Suresh Srinivas edited comment on HDFS-2301 at 9/28/11 11:23 PM:
-----------------------------------------------------------------

h3. Current behavior:
Namenode starts the following services:
# FSNamesystem secret manager
# FSNamesystem activation performs:
#* NNResourceChecker, activate block manager, starts lease manager thread, resource monitor
threads and register mbeans and initializes metrics.
# Starts http server, rpc server, trash emptier, service plugins

Backup node starts all the services above except:
# secret manager
# Sets lease period to MAX_VALUE, effectively not use lease manager.



h3. Proposed behavior:
Following services are started as common to both active and standby namenode states:
# Starts common FSNamesystem common services to both active and standby
#* NNResourceChecker, activate block manager, resource monitor threads and register mbeans
and initializes metrics.
# Starts http server, rpc server, service plugins

Following services are started when moving to active state and stopped when exiting active
state:
# Start/stops FSNamesystem active services:
#* Which starts/stop FSNamesystem secrtet manager and lease manager.

Currently standby state services are marked as TODO. This is where reading editlogs from active
will be handled.

BackupNode with this patch will no longer start lease manager thread.

h3. Code changes
# HAState no longer knows about NameNode. Instead NameNode exposes HAContext
# Reorganized and cleaned up the code for starting and stopping services.


                
      was (Author: sureshms):
    h3. Current behavior:
Namenode starts the following services:
# FSNamesystem secret manager
# FSNamesystem activation performs:
#* NNResourceChecker, activate block manager, starts lease manager thread, resource monitor
threads and register mbeans and initializes metrics.
# Starts http server, rpc server, trash emptier, service plugins

Backup node starts all the services above except:
# secret manager
# Sets lease period to MAX_VALUE, effectively not use lease manager.



h3. Proposed behavior:
Following services are started as common to both active and standby namenode states:
# Starts common FSNamesystem common services to both active and standby
#* TODO
# Starts http server, rpc server, service plugins

Following services are started when moving to active state and stopped when exiting active
state:
# Start/stops FSNamesystem active services:
#* Which starts/stop FSNamesystem secrtet manager and lease manager.

Currently standby state services are marked as TODO. This is where reading editlogs from active
will be handled.

BackupNode with this patch will no longer start lease manager thread.

h3. Code changes
# HAState no longer knows about NameNode. Instead NameNode exposes HAContext
# Reorganized and cleaned up the code for starting and stopping services.

                  
> Start/stop appropriate namenode internal services during transition to active and standby
> -----------------------------------------------------------------------------------------
>
>                 Key: HDFS-2301
>                 URL: https://issues.apache.org/jira/browse/HDFS-2301
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: name-node
>    Affects Versions: HA branch (HDFS-1623)
>            Reporter: Suresh Srinivas
>            Assignee: Suresh Srinivas
>             Fix For: HA branch (HDFS-1623)
>
>         Attachments: HDFS-2301.txt
>
>
> These changes are related to HDFS-1974 which introduced active and standby states. This
jira will address starting and stopping appropriate NN services when entering and existing
active and standby states.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message