elek commented on pull request #2055:
URL: https://github.com/apache/ozone/pull/2055#issuecomment-803952829
> We also discussed yesterday that, let's not make index as the property to
chunkInfo. I am more comfortable if we make that as property to
ContainerBlockID instead. Thanks for removing them in latest patch.
Sure, I removed it, but during my git rebase it's not fully moved (I manage
10-12 patches today and trying to sync them one-by-one with different PRs)
Now it should be good:
```
/**
* Block ID that uniquely identify a block in Datanode.
*/
message DatanodeBlockID {
required int64 containerID = 1;
required int64 localID = 2;
optional uint64 blockCommitSequenceId = 3 [default = 0];
optional int32 replicaIndex = 4;
}
```
> Please add the replication index to org.apache.hadoop.hdds.client.BlockID
and org.apache.hadoop.hdds.client.ContainerBlockID
> This will be needed when we create container from client, we anyway pass
blockID to DN. If index available there, then DN know which index.
`org.apache.hadoop.hdds.client.ContainerBlockID` doesn't need any
modification as it contains the generic id (which points to the replication
set) and pipelines (where each element can have replicaIndex information)
Same is true for `BlockID` some cases we need this information in some cases
we don't I propose to add the information only when it's required. We will see
it in the next patches with adding the client side / datanode side changes, and
we can extend `BlockID` and create a `BlockInstanceId` when required...
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
