[
https://issues.apache.org/jira/browse/HDFS-15443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17529787#comment-17529787
]
Brahma Reddy Battula commented on HDFS-15443:
---------------------------------------------
Basically this config is to specify the maximum number of files that a DataNode
can serve at any one time. Yes, CPU ( for checksum calculations) ,number of
"open file descriptors" and memory availability for the stack's (Even based on
the network bandwidth).
{quote}are at 20k today and have 200TB of usable space each server, want to see
if we can increase to 32k.
{quote}
if you see, there are above mentioned resources are enough and you are not
reaching the expected the throughput you can increase. But IMO,20k itself is
big number.
> Setting dfs.datanode.max.transfer.threads to a very small value can cause
> strange failure.
> ------------------------------------------------------------------------------------------
>
> Key: HDFS-15443
> URL: https://issues.apache.org/jira/browse/HDFS-15443
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode
> Reporter: AMC-team
> Assignee: AMC-team
> Priority: Major
> Fix For: 3.3.1, 3.4.0
>
> Attachments: HDFS-15443.000.patch, HDFS-15443.001.patch,
> HDFS-15443.002.patch, HDFS-15443.003.patch
>
>
> Configuration parameter dfs.datanode.max.transfer.threads is to specify the
> maximum number of threads to use for transferring data in and out of the DN.
> This is a vital param that need to tune carefully.
> {code:java}
> // DataXceiverServer.java
> // Make sure the xceiver count is not exceeded
> intcurXceiverCount = datanode.getXceiverCount();
> if (curXceiverCount > maxXceiverCount) {
> thrownewIOException("Xceiver count " + curXceiverCount
> + " exceeds the limit of concurrent xceivers: "
> + maxXceiverCount);
> }
> {code}
> There are many issues that caused by not setting this param to an appropriate
> value. However, there is no any check code to restrict the parameter.
> Although having a hard-and-fast rule is difficult because we need to consider
> number of cores, main memory etc, *we can prevent users from setting this
> value to an absolute wrong value by accident.* (e.g. a negative value that
> totally break the availability of datanode.)
> *How to fix:*
> Add proper check code for the parameter.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
