You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "Eli Collins (Created) (JIRA)" <ji...@apache.org> on 2011/12/30 23:12:30 UTC
[jira] [Created] (HDFS-2736) HA: support separate SBN and 2NN?
HA: support separate SBN and 2NN?
---------------------------------
Key: HDFS-2736
URL: https://issues.apache.org/jira/browse/HDFS-2736
Project: Hadoop HDFS
Issue Type: Sub-task
Components: ha
Affects Versions: HA branch (HDFS-1623)
Reporter: Eli Collins
HDFS-2291 adds support for making the SBN capable of checkpointing, seems like we may also need to support the 2NN checkpointing as well. Eg if we fail over to the SBN does it continue to checkpoint? If not the log grows unbounded until the old primary comes back, if so does that create performance problems since the primary wasn't previously checkpointing?
--
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
[jira] [Resolved] (HDFS-2736) HA: support separate SBN and 2NN?
Posted by "Eli Collins (Resolved) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-2736?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Eli Collins resolved HDFS-2736.
-------------------------------
Resolution: Won't Fix
I was thinking about the case where the former active remains dead for some time. But this case is problematic for a number of other reasons (eg with just a single host we can't failback if necessary) so I think its reasonable to require users start another SBN instead also deploying a 2NN. Closing as won't fix, can re-open if others disagree.
> HA: support separate SBN and 2NN?
> ---------------------------------
>
> Key: HDFS-2736
> URL: https://issues.apache.org/jira/browse/HDFS-2736
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ha
> Affects Versions: HA branch (HDFS-1623)
> Reporter: Eli Collins
>
> HDFS-2291 adds support for making the SBN capable of checkpointing, seems like we may also need to support the 2NN checkpointing as well. Eg if we fail over to the SBN does it continue to checkpoint? If not the log grows unbounded until the old primary comes back, if so does that create performance problems since the primary wasn't previously checkpointing?
--
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
[jira] [Commented] (HDFS-2736) HA: support separate SBN and 2NN?
Posted by "Todd Lipcon (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-2736?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177791#comment-13177791 ]
Todd Lipcon commented on HDFS-2736:
-----------------------------------
bq. if we fail over to the SBN does it continue to checkpoint?
If the only active NN comes back, it will start in standby mode and then do checkpoints against the other one. It's only if the former active remains dead for some time that we might have a problem.
bq. If not the log grows unbounded until the old primary comes back, if so does that create performance problems since the primary wasn't previously checkpointing?
Certainly the first startup after a long outage will be slower, since it has to replay a lot of transactions. But it's not any different than what happens today if the 2NN goes down for some length of time.
Given that people currently run successfully with only a single 2NN, I don't think this is particularly high priority. Do you disagree?
> HA: support separate SBN and 2NN?
> ---------------------------------
>
> Key: HDFS-2736
> URL: https://issues.apache.org/jira/browse/HDFS-2736
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ha
> Affects Versions: HA branch (HDFS-1623)
> Reporter: Eli Collins
>
> HDFS-2291 adds support for making the SBN capable of checkpointing, seems like we may also need to support the 2NN checkpointing as well. Eg if we fail over to the SBN does it continue to checkpoint? If not the log grows unbounded until the old primary comes back, if so does that create performance problems since the primary wasn't previously checkpointing?
--
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
[jira] [Updated] (HDFS-2736) HA: support 2NN with SBN
Posted by "Eli Collins (Updated) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-2736?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Eli Collins updated HDFS-2736:
------------------------------
Summary: HA: support 2NN with SBN (was: HA: support separate SBN and 2NN?)
> HA: support 2NN with SBN
> ------------------------
>
> Key: HDFS-2736
> URL: https://issues.apache.org/jira/browse/HDFS-2736
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ha
> Affects Versions: HA branch (HDFS-1623)
> Reporter: Eli Collins
>
> HDFS-2291 adds support for making the SBN capable of checkpointing, seems like we may also need to support the 2NN checkpointing as well. Eg if we fail over to the SBN does it continue to checkpoint? If not the log grows unbounded until the old primary comes back, if so does that create performance problems since the primary wasn't previously checkpointing?
--
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
[jira] [Commented] (HDFS-2736) HA: support 2NN with SBN
Posted by "Todd Lipcon (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-2736?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177858#comment-13177858 ]
Todd Lipcon commented on HDFS-2736:
-----------------------------------
Agreed. Currently we also basically support multiple-standby. Even though it wasn't a goal of the implementation, it basically fell out for free.
> HA: support 2NN with SBN
> ------------------------
>
> Key: HDFS-2736
> URL: https://issues.apache.org/jira/browse/HDFS-2736
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ha
> Affects Versions: HA branch (HDFS-1623)
> Reporter: Eli Collins
>
> HDFS-2291 adds support for making the SBN capable of checkpointing, seems like we may also need to support the 2NN checkpointing as well. Eg if we fail over to the SBN does it continue to checkpoint? If not the log grows unbounded until the old primary comes back, if so does that create performance problems since the primary wasn't previously checkpointing?
--
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