[
https://issues.apache.org/jira/browse/HDFS-13279?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16427301#comment-16427301
]
Ajay Kumar commented on HDFS-13279:
-----------------------------------
[~Tao Jie], Thanks for updating the patch. Few comments
* We should try to avoid any change in default behaviour.
** Don't remove the default impl entry from core-default.xml.
** Similarly we can do this with almost no change in Network Topology,
DFSNetworkTopology and TestBalancerWithNodeGroup. We can update
DatanodeManager#init to instantiate according to value of "net.topology.impl".
* NetworkTopologyWithWeightedRack
** Shall we override chooseRandom & chooseRandomWithStorageType to select
next random node utilizing weights of rack. i.e first select rack according to
weighted probability and then select a node in that rack.
*** choose rack according to weighted probability and excluded nodes.
*** choose random node from rack selected on first step.
*** repeat until a node is found or all racks in scope are tried at least once.
** L29, shall we simplify javadoc for class to something like "The class
extends NetworkTopology to represents racks weighted according to number of
datanodes."
** L38 LimitedPrivate to Private
* WeightedRackBlockPlacementPolicy
** L43, chooseDataNode: Instead of choosing datanode twice we can just call
super.chooseRandom if we overide chooseRandom in
NetworkTopologyWithWeightedRack. This way we can avoid calling chooseRandom
twice.
** Private and Unstable annotation for class
> Datanodes usage is imbalanced if number of nodes per rack is not equal
> ----------------------------------------------------------------------
>
> Key: HDFS-13279
> URL: https://issues.apache.org/jira/browse/HDFS-13279
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 2.8.3, 3.0.0
> Reporter: Tao Jie
> Assignee: Tao Jie
> Priority: Major
> Attachments: HDFS-13279.001.patch, HDFS-13279.002.patch,
> HDFS-13279.003.patch, HDFS-13279.004.patch, HDFS-13279.005.patch
>
>
> In a Hadoop cluster, number of nodes on a rack could be different. For
> example, we have 50 Datanodes in all and 15 datanodes per rack, it would
> remain 5 nodes on the last rack. In this situation, we find that storage
> usage on the last 5 nodes would be much higher than other nodes.
> With the default blockplacement policy, for each block, the first
> replication has the same probability to write to each datanode, but the
> probability for the 2nd/3rd replication to write to the last 5 nodes would
> much higher than to other nodes.
> Consider we write 50 blocks to such 50 datanodes. The first rep of 100 block
> would distirbuted to 50 node equally. The 2rd rep of blocks which the 1st rep
> is on rack1(15 reps) would send equally to other 35 nodes and each nodes
> receive 0.428 rep. So does blocks on rack2 and rack3. As a result, node on
> rack4(5 nodes) would receive 1.29 replications in all, while other node would
> receive 0.97 reps.
> ||-||Rack1(15 nodes)||Rack2(15 nodes)||Rack3(15 nodes)||Rack4(5 nodes)||
> |From rack1|-|15/35=0.43|0.43|0.43|
> |From rack2|0.43|-|0.43|0.43|
> |From rack3|0.43|0.43|-|0.43|
> |From rack4|5/45=0.11|0.11|0.11|-|
> |Total|0.97|0.97|0.97|1.29|
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
