[
https://issues.apache.org/jira/browse/HDFS-1108?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13090001#comment-13090001
]
Suresh Srinivas commented on HDFS-1108:
---------------------------------------
bq. Data loss on failover is not acceptable. I thought that all transactions
will be passed to the standby node and processed in RAM there, but not synced
to disk. On failover you will have completely up to date namespace in SBN. No
need to sync.
If BackupNode is the only node that has the more current editlog, that means
any restarts without involving the BackupNode could cause consistency issue.
Isn't BackupNode stream treated as editlog. How do you differentiate between
logSync() to that stream Vs stream to the disk?
My vote is to call logSync() after block allocation. While this might hold a
thread a little longer, I do not see an issue, unless a benchmark shows serious
regression. Lets not optimize prematurely if this is not an issue.
bq. As I understood Suresh is trying to build a universal HA framework
applicable to both.
Again I have explained this in several jiras and in our conversations. It is
not "Universal HA framework". The approach 2 has different requirements that
many are interested in. I have repeated this many times.
> Log newly allocated blocks
> --------------------------
>
> Key: HDFS-1108
> URL: https://issues.apache.org/jira/browse/HDFS-1108
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: name-node
> Reporter: dhruba borthakur
> Assignee: Todd Lipcon
> Fix For: HA branch (HDFS-1623)
>
> Attachments: HDFS-1108.patch, hdfs-1108-habranch.txt, hdfs-1108.txt
>
>
> The current HDFS design says that newly allocated blocks for a file are not
> persisted in the NN transaction log when the block is allocated. Instead, a
> hflush() or a close() on the file persists the blocks into the transaction
> log. It would be nice if we can immediately persist newly allocated blocks
> (as soon as they are allocated) for specific files.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
