[
https://issues.apache.org/jira/browse/HDFS-724?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hairong Kuang updated HDFS-724:
-------------------------------
Attachment: pipelineHeartbeat.patch
OK, here comes a patch that does the following:
1. when a pipeline is idle for half of the read timeout, dfsclient sends a
heartbeat packet that has a sequence number of -1 and a data length of 0;
2. At the client side, a heartbeat packet does not get put the data queue and
hence not in the ack queue;
3. At the datanode side, the block receiver treats a heartbeat packet mostly
like a regular data packet; A heatbeat packet is put in the ack queue and its
ack is also the same as that of a regular data packet;
4. The ack to a heartbeat packet may also indicate failures in the pipeline and
therefore triggers pipeline recovery.
> Pipeline close hangs if one of the datanode is not responsive.
> --------------------------------------------------------------
>
> Key: HDFS-724
> URL: https://issues.apache.org/jira/browse/HDFS-724
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: data-node, hdfs client
> Affects Versions: 0.21.0
> Reporter: Tsz Wo (Nicholas), SZE
> Assignee: Hairong Kuang
> Priority: Blocker
> Fix For: 0.21.0
>
> Attachments: h724_20091021.patch, pipelineHeartbeat.patch
>
>
> In the new pipeline design, pipeline close is implemented by sending an
> additional empty packet. If one of the datanode does not response to this
> empty packet, the pipeline hangs. It seems that there is no timeout.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
