[
https://issues.apache.org/jira/browse/HADOOP-2065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12595777#action_12595777
]
Konstantin Shvachko commented on HADOOP-2065:
---------------------------------------------
This looks correct.
I have a concern that we probably need to move this into a separate
data-structure. So that not to refactor it later as we did with other block
collections. So,
- I'd highly recommend to create a separate class say CorruptedReplicaMap or
CorruptedReplicas or CorruptedBlocks.
The important thing that all members and methods related to corrupted replicas
currently populated in FSNamesystem like invalidateCorruptReplicas(),
markBlockAsCorrupt() should belong to this class as well as add(), get(),
remove(), isCorrupt().
This is simialr to PendingReplicationBlocks, BlocksMap and
UnderReplicatedBlocks.
- The method of the new class should clearly distinguish between Block and
BlockInfo parameters. I presume most parameters will be BlockInfo, because this
collection holds only those blocks that belong to the BlocksMap.
- processPendingReplications()
if (num.liveReplicas == 0) continue; is not necessary, because
neededReplications.add() does that inside.
This code should just be removed.
> Replication policy for corrupted block
> ---------------------------------------
>
> Key: HADOOP-2065
> URL: https://issues.apache.org/jira/browse/HADOOP-2065
> Project: Hadoop Core
> Issue Type: Bug
> Components: dfs
> Affects Versions: 0.14.1
> Reporter: Koji Noguchi
> Assignee: lohit vijayarenu
> Fix For: 0.18.0
>
> Attachments: HADOOP-2065-2.patch, HADOOP-2065-3.patch,
> HADOOP-2065-4.patch, HADOOP-2065-5.patch, HADOOP-2065-6.patch,
> HADOOP-2065.patch
>
>
> Thanks to HADOOP-1955, even if one of the replica is corrupted, the block
> should get replicated from a good replica relatively fast.
> Created this ticket to continue the discussion from
> http://issues.apache.org/jira/browse/HADOOP-1955#action_12531162.
> bq. 2. Delete corrupted source replica
> bq. 3. If all replicas are corrupt, stop replication.
> For (2), it'll be nice if the namenode can delete the corrupted block if
> there's a good replica on other nodes.
> For (3), I prefer if the namenode can still replicate the block.
> Before 0.14, if the file was corrupted, users were still able to pull the
> data and decide if they want to delete those files. (HADOOP-2063)
> In 0.14 and later, we cannot/don't replicate these blocks so they eventually
> get lost.
> To make the matters worse, if the corrupted file is accessed, all the
> corrupted replicas would be deleted except for one and stay as replication
> factor of 1 forever.
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
