[
https://issues.apache.org/jira/browse/HADOOP-9640?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13990308#comment-13990308
]
Ming Ma commented on HADOOP-9640:
---------------------------------
Thanks, Chris.
1. The current approach drops call when RPC queue is full and the client relies
on RPC timeout. It will be interesting to confirm if it is useful to have RPC
server throw some exception back to client and have client do exponential back
off; or maybe just block the RPC reader thread instead.
2. RPC-based approach didn't account for http request such as webHDFS. Based on
some test results, it seems Jetty uses around 250 threads, small compared to
the thousands of RPC handler threads. a) The bad application traffic from
webHDFS still has impact on RPC latency, not as severe compared to the RPC
case. b), if there are SLA jobs based on webHDFS, then the RPC throttling won't
help much.
> RPC Congestion Control with FairCallQueue
> -----------------------------------------
>
> Key: HADOOP-9640
> URL: https://issues.apache.org/jira/browse/HADOOP-9640
> Project: Hadoop Common
> Issue Type: Improvement
> Affects Versions: 3.0.0, 2.2.0
> Reporter: Xiaobo Peng
> Assignee: Chris Li
> Labels: hdfs, qos, rpc
> Attachments: FairCallQueue-PerformanceOnCluster.pdf,
> MinorityMajorityPerformance.pdf, NN-denial-of-service-updated-plan.pdf,
> faircallqueue.patch, faircallqueue2.patch, faircallqueue3.patch,
> faircallqueue4.patch, faircallqueue5.patch, faircallqueue6.patch,
> faircallqueue7_with_runtime_swapping.patch,
> rpc-congestion-control-draft-plan.pdf
>
>
> Several production Hadoop cluster incidents occurred where the Namenode was
> overloaded and failed to respond.
> We can improve quality of service for users during namenode peak loads by
> replacing the FIFO call queue with a [Fair Call
> Queue|https://issues.apache.org/jira/secure/attachment/12616864/NN-denial-of-service-updated-plan.pdf].
> (this plan supersedes rpc-congestion-control-draft-plan).
> Excerpted from the communication of one incident, “The map task of a user was
> creating huge number of small files in the user directory. Due to the heavy
> load on NN, the JT also was unable to communicate with NN...The cluster
> became responsive only once the job was killed.”
> Excerpted from the communication of another incident, “Namenode was
> overloaded by GetBlockLocation requests (Correction: should be getFileInfo
> requests. the job had a bug that called getFileInfo for a nonexistent file in
> an endless loop). All other requests to namenode were also affected by this
> and hence all jobs slowed down. Cluster almost came to a grinding
> halt…Eventually killed jobtracker to kill all jobs that are running.”
> Excerpted from HDFS-945, “We've seen defective applications cause havoc on
> the NameNode, for e.g. by doing 100k+ 'listStatus' on very large directories
> (60k files) etc.”
--
This message was sent by Atlassian JIRA
(v6.2#6252)
