You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Manikandan R (JIRA)" <ji...@apache.org> on 2018/11/28 17:52:00 UTC
[jira] [Comment Edited] (YARN-6523) Newly retrieved security Tokens
are sent as part of each heartbeat to each node from RM which is not
desirable in large cluster
[ https://issues.apache.org/jira/browse/YARN-6523?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16702196#comment-16702196 ]
Manikandan R edited comment on YARN-6523 at 11/28/18 5:51 PM:
--------------------------------------------------------------
Thanks for very informative comments.
{quote}All we need to do here is cache the list of SystemCredentialsForAppsProto values and have NodeHearbeatResponse take that list of protos rather than a Map<String,ByteBuffer> for the system credentials. NodeHeartbeatResponsePBImpl can then call addAllSystemCredentialsForApps on the builder when it builds its protocol buffer{quote}
Taken care and modified related unit test cases as well.
{quote}TestYarnServerApiClasses#testNodeHeartbeatResponsePBImpl has two "// create token2" comments, and I'm assuming only one of them is accurate.{quote}
Correct. Cleaned it up.
{quote}The very long unit test was removed but equivalent tests were not added.{quote}
Sorry, I missed to include this change in earlier patch. As I said earlier, configuring appropriate token expiry time reduce overall test case execution time. Now it is between 10-13 secs and hope it is fine. However, I am open to changes.
Also taken care of other minor nits - checkstyle warnings, whitespace issues, debug stmts etc.
was (Author: manirajv06@gmail.com):
Thanks for very informative comments.
{quote}All we need to do here is cache the list of SystemCredentialsForAppsProto values and have NodeHearbeatResponse take that list of protos rather than a Map<String,ByteBuffer> for the system credentials. NodeHeartbeatResponsePBImpl can then call addAllSystemCredentialsForApps on the builder when it builds its protocol buffer\{quote}
Taken care and modified related unit test cases as well.
{quote}TestYarnServerApiClasses#testNodeHeartbeatResponsePBImpl has two "// create token2" comments, and I'm assuming only one of them is accurate.\{quote}
Correct.
{quote}The very long unit test was removed but equivalent tests were not added.\{quote}
Sorry, I missed to include this change in earlier patch. As I said earlier, configuring appropriate token expiry time reduced overall test case execution time. Now it is between 10-13 secs and hope it is fine. However, I am open to changes.
Also taken care of other minor nits - checkstyle warnings, whitespace issues, debug stmts etc.
> Newly retrieved security Tokens are sent as part of each heartbeat to each node from RM which is not desirable in large cluster
> -------------------------------------------------------------------------------------------------------------------------------
>
> Key: YARN-6523
> URL: https://issues.apache.org/jira/browse/YARN-6523
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: RM
> Affects Versions: 2.8.0, 2.7.3
> Reporter: Naganarasimha G R
> Assignee: Manikandan R
> Priority: Major
> Attachments: YARN-6523.001.patch, YARN-6523.002.patch, YARN-6523.003.patch, YARN-6523.004.patch, YARN-6523.005.patch, YARN-6523.006.patch, YARN-6523.007.patch
>
>
> Currently as part of heartbeat response RM sets all application's tokens though all applications might not be active on the node. On top of it NodeHeartbeatResponsePBImpl converts tokens for each app into SystemCredentialsForAppsProto. Hence for each node and each heartbeat too many SystemCredentialsForAppsProto objects were getting created.
> We hit a OOM while testing for 2000 concurrent apps on 500 nodes cluster with 8GB RAM configured for RM
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org