hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Josh Elser <els...@apache.org>
Subject Re: [DISCUSS] More Shading
Date Wed, 12 Apr 2017 15:28:41 GMT

Sean Busbey wrote:
> On Tue, Apr 11, 2017 at 11:43 PM Nick Dimiduk<ndimiduk@gmail.com>  wrote:
>>> This effort is about our internals. We have a mess of other components
>> all
>>> up inside us such as HDFS, etc., each with their own sets of dependencies
>>> many of which we have in common. This project t is about making it so we
>>> can upgrade at a rate independent of when our upstreamers choose to
>> change.
>> Pardon as I try to get a handle on the intention behind this thread.
>> If the above quote is true, then I think what we want is a set of shaded
>> Hadoop client libs that we can depend on so as to not get all the
>> transitive deps. Hadoop doesn't provide it, but we could do so ourselves
>> with (yet another) module in our project. Assuming, that is, the upstream
>> client interfaces are well defined and don't leak stuff we care about. It
>> also creates a terrible nightmare for anyone downstream of us who
>> repackages HBase. The whole thing is extremely error-prone, because there's
>> not very good tooling for this. Realistically, we end up with a combination
>> of the enforcer plugin and maybe our own custom plugin to ensure clean
>> transitive dependencies...
> Hadoop does provide a shaded client as of the 3.0.0* release line. We could
> push as a community for a version of that for Hadoop's branch-2.
> Unfortunately, that shaded client won't help where we're reaching into the
> guts of Hadoop (like our reliance on their web stuff).

Well put, Nick.

With Sean's point about the Hadoop shaded client, it seems to me that we 
have things which could be pursued in parallel:

1) Roadmap to Hadoop3 (and shaded hdfs client).
2) Identify components which we use from Hadoop, for each component:
   2a) Work with Hadoop to isolate that component from other cruft (best 
example is the Configuration class -- you get something like 8MB of 
"jar" just to parse an xml file).
   2b) Pull the implementation into HBase, removing dependency from 
Hadoop entirely.

I think that both of these can/should be done in parallel to the 
isolation of the dependencies which HBase requires (isolating ourselves 
from upstream, and isolating downstream from us).

View raw message