[
https://issues.apache.org/jira/browse/HDFS-14090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16863761#comment-16863761
]
Yiqun Lin commented on HDFS-14090:
----------------------------------
Hi [~crh], I also prefer the approach one, but some comments from me.
{quote}As mentioned, if a downstream namenode is slow/busy enough that permits
are unavailable, routers would throw StandByException exception to the client.
This would in turn trigger a failover behavior at the client side and clients
would connect to a different router in the cluster.
{quote}
I'm not sure if this was implemented now. From my impression, the client side
only connect to the one specific Router. And the StandbyException will trigger
the router client to failover to another NN not another Router.
Even though the client side can failover to another Router, the slowness issue
can still exist in downstream NN. That is say, the failover behavior across
Routers is meaningless to improve the slowness in downstream NN when NN is
busy. The new Router will also quickly reach the permit limitation when NN is
in busy state. So the correct behavior maybe that throw a normal IOException
like Fair Call Queue bakoff logic did.
> RBF: Improved isolation for downstream name nodes.
> --------------------------------------------------
>
> Key: HDFS-14090
> URL: https://issues.apache.org/jira/browse/HDFS-14090
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: CR Hota
> Assignee: CR Hota
> Priority: Major
> Attachments: HDFS-14090-HDFS-13891.001.patch,
> HDFS-14090-HDFS-13891.002.patch, HDFS-14090-HDFS-13891.003.patch,
> HDFS-14090-HDFS-13891.004.patch, HDFS-14090-HDFS-13891.005.patch, RBF_
> Isolation design.pdf
>
>
> Router is a gateway to underlying name nodes. Gateway architectures, should
> help minimize impact of clients connecting to healthy clusters vs unhealthy
> clusters.
> For example - If there are 2 name nodes downstream, and one of them is
> heavily loaded with calls spiking rpc queue times, due to back pressure the
> same with start reflecting on the router. As a result of this, clients
> connecting to healthy/faster name nodes will also slow down as same rpc queue
> is maintained for all calls at the router layer. Essentially the same IPC
> thread pool is used by router to connect to all name nodes.
> Currently router uses one single rpc queue for all calls. Lets discuss how we
> can change the architecture and add some throttling logic for
> unhealthy/slow/overloaded name nodes.
> One way could be to read from current call queue, immediately identify
> downstream name node and maintain a separate queue for each underlying name
> node. Another simpler way is to maintain some sort of rate limiter configured
> for each name node and let routers drop/reject/send error requests after
> certain threshold.
> This won’t be a simple change as router’s ‘Server’ layer would need redesign
> and implementation. Currently this layer is the same as name node.
> Opening this ticket to discuss, design and implement this feature.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
