[
https://issues.apache.org/jira/browse/HADOOP-4584?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12676391#action_12676391
]
Konstantin Shvachko commented on HADOOP-4584:
---------------------------------------------
So what is wrong with just going with the original proposal in this jira that
is: prepare block reports in a separate thread without delaying heartbeats and
other commands, and sending them as soon as they are ready by
{{offerService()}}. This seem to be the mission declared by the issue, and
changing block reports to be memory based is an add-on, which is not required
to solve the problem stated.
I understand Dhruba's concerns about reliability. I can add to this that memory
based reports can also slow down cleaning up disks from unnecessary blocks,
which may be critical if the data-node is close to running out of disk space.
My approach would be to drop the in-memory block report part and commit the
rest. The in-memory reports can be discussed in a subsequent issue.
I think that would be enough of a change by itself, because there may be a
dangerous race condition between {{blockReceived()}} and {{blockReport()}} if
it is not done right, as we had seen before.
> Slow generation of blockReport at DataNode causes delay of sending heartbeat
> to NameNode
> ----------------------------------------------------------------------------------------
>
> Key: HADOOP-4584
> URL: https://issues.apache.org/jira/browse/HADOOP-4584
> Project: Hadoop Core
> Issue Type: Bug
> Components: dfs
> Reporter: Hairong Kuang
> Assignee: Suresh Srinivas
> Fix For: 0.20.0
>
> Attachments: 4584.patch, 4584.patch, 4584.patch, 4584.patch,
> 4584.patch, 4584.patch
>
>
> sometimes due to disk or some other problems, datanode takes minutes or tens
> of minutes to generate a block report. It causes the datanode not able to
> send heartbeat to NameNode every 3 seconds. In the worst case, it makes
> NameNode to detect a lost heartbeat and wrongly decide that the datanode is
> dead.
> It would be nice to have two threads instead. One thread is for scanning data
> directories and generating block report, and executes the requests sent by
> NameNode; Another thread is for sending heartbeats, block reports, and
> picking up the requests from NameNode. By having these two threads, the
> sending of heartbeats will not get delayed by any slow block report or slow
> execution of NameNode requests.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
