Anu Engineer commented on HDFS-11010:

bq.  besides when we asked Google to rehost the PB 2.5 binary when they 
migrated to github from google code.

That is exactly my concern. If we are going to take a dependency on an 
unsupported component, we must have a clear plan how we are going to do that. 
As far as google is concerned they don't really maintain 2.5, depending on a 
binary only component seems like a bad idea, irrespective of how stable that 
component is.

I am glad we are having this discussion, I seems like what you are proposing is 
the option 4 that I have suggested -- Be forever stuck in protobuf 2 or at 
least till the lifetime of Hadoop 3. I have no issues if that is the option 
that we want to pursue, but I do hope we have a better way of handling this 
orphaned component. Already google has indicated that they are not supporting 
this and has been extremely gracious by accommodating us. But for how long ? I 
am concerned that we might be overstaying our welcome.

bq. Are there any known bugs or issues with PB2.5 that are addressed by PB3?
No, but the fact that there is no one maintaining it (including us) is the 
concern that I am trying to articulate. So I am not worried about bugs per se, 
this is more about having the right forward looking strategy. Pushing it into 
Hadoop 4 is a way of not quantifying the risk we have; The reason why  I am 
proposing a move today is because in hadoop 3 we can depreciate protobuf 2 and 
in hadoop 4 then can actually remove  protobuf 2.  

> Update to Protobuf 3 in trunk
> -----------------------------
>                 Key: HDFS-11010
>                 URL: https://issues.apache.org/jira/browse/HDFS-11010
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: ipc
>    Affects Versions: 3.0.0-alpha2
>            Reporter: Anu Engineer
> This JIRA is to propose that we should move to protobuf 3 from protobuf 2.5. 
> This allows us to stay with the newer versions of the protobuf.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org

Reply via email to