[
https://issues.apache.org/jira/browse/HADOOP-5189?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12671794#action_12671794
]
Luca Telloli commented on HADOOP-5189:
--------------------------------------
Jeff, I think that for being a logging abstraction, BookKeeper does not
implement it but provides a way to achieve it. That is, if the NameNode dies,
another process could reconstruct the fsimage starting from a set of ledgers.
Actually, any number of independent processes could do it, as soon as they read
from the same set of ledgers. This process could definitely be the
SecondaryNameNode.
I guess that such a process could be implemented to allow warm standby, as
proposed in the description. First thought: ZooKeeper could be used to store
such an information (the set of ledger IDs that belong to a single fsimage) and
the secondary NameNodes could periodically poll for changes (or maybe using
some notification mechanism).
> Integration with BookKeeper logging system
> ------------------------------------------
>
> Key: HADOOP-5189
> URL: https://issues.apache.org/jira/browse/HADOOP-5189
> Project: Hadoop Core
> Issue Type: New Feature
> Affects Versions: 0.19.0
> Reporter: Luca Telloli
> Attachments: create.png, HADOOP-5189.patch
>
>
> BookKeeper is a system to reliably log streams of records
> (https://issues.apache.org/jira/browse/ZOOKEEPER-276). The NameNode is a
> natural target for such a system for being the metadata repository of the
> entire file system for HDFS.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
