Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-02-09 Thread Haohui Mai
The HDFS-5698 branch has been merged to trunk. We plan to continue to test the feature on bigger clusters and to merge it to branch-2 next week. Thanks, Haohui On Fri, Feb 7, 2014 at 3:59 PM, Haohui Mai wrote: > With 3 binding +1 and 2 non-binding +1 votes, the vote has passed. > > The code wi

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-02-07 Thread Haohui Mai
With 3 binding +1 and 2 non-binding +1 votes, the vote has passed. The code will be merged soon (pending Jenkins). We will address any remaining issues in trunk. Thanks to everyone who voted and provided their feedback. Regards, Haohui On Wed, Feb 5, 2014 at 12:47 PM, Chris Nauroth wrote: > +

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-02-05 Thread Chris Nauroth
+1 Thank you, Haohui Mai and Jing Zhao. I'm definitely looking forward to fewer metadata upgrades in the future. :-) Chris Nauroth Hortonworks http://hortonworks.com/ On Mon, Feb 3, 2014 at 10:49 PM, Akira AJISAKA wrote: > +1 (non-binding). > > This feature is very good for the developers t

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-02-03 Thread Akira AJISAKA
+1 (non-binding). This feature is very good for the developers to reduce the tasks for supporting FSImage compatibility. Thanks, Akira (2014/01/31 7:37), Haohui Mai wrote: Hello all, I would like to call a vote to merge of the new protobuf-based FSImage into trunk. The changes introduce a pr

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-01-31 Thread Suresh Srinivas
+1. This is an important feature that gets rid of all the custom serialization and unnecessary code around it. Helps in rolling upgrade scenarios too! Nice job Haohui Mai and Jing Zhao! On Thu, Jan 30, 2014 at 2:37 PM, Haohui Mai wrote: > Hello all, > > I would like to call a vote to merge of

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-01-30 Thread Azuryy Yu
+1 This was very important improvement. hope it release at 2.4 On Jan 31, 2014 6:37 AM, "Haohui Mai" wrote: > Hello all, > > I would like to call a vote to merge of the new protobuf-based FSImage into > trunk. > > The changes introduce a protobuf-based FSImage format into HDFS. It > separates

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-01-30 Thread Haohui Mai
Hi Todd, This is a great question. I don't see a lot of work from the implementation prospective. I anticipate that the branch-2 patch will be almost identical to the one for trunk. We plan on merging it into branch-2 after more testings. In my opinion it should fit in the 2.4 timeline. And to c

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-01-30 Thread Todd Lipcon
One question - what's the intention with this work with regard to branch-2? Do you plan to merge it to branch-2, and if so, are you thinking for 2.4 or later? I'm a little afraid of the divergence between trunk and branch-2 in this area, since we'll likely want to add new features which require se

Re: [VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-01-30 Thread Jing Zhao
+1. I think this is a very useful improvement. In the past when I wanted to add a new snapshot field which needs to be persisted in fsimage, I found that I would break the binary compatibility thus some workaround had to be used (see HDFS-5428). Now protobuf gives us this flexibility to make these

[VOTE] Merge HDFS-5698 Use protobuf to serialize / deserialize FSImage to trunk

2014-01-30 Thread Haohui Mai
Hello all, I would like to call a vote to merge of the new protobuf-based FSImage into trunk. The changes introduce a protobuf-based FSImage format into HDFS. It separates the responsibility of serializing and reconstructing the namespace in the NameNode using Google's Protocol Buffers (protobuf)