Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HDFS >> mail # dev >> VOTE: HDFS-347 merge


Copy link to this message
-
Re: VOTE: HDFS-347 merge
On Mon, Feb 25, 2013 at 1:50 PM, Chris Douglas <[EMAIL PROTECTED]> wrote:
> On Mon, Feb 25, 2013 at 1:16 PM, Eli Collins <[EMAIL PROTECTED]> wrote:
>> I think we need a transition period when any kinks are worked out of
>> 347 but I don't think we need one alpha/beta release where both
>> mechanisms are supported (because 2246 was just a short term solution
>> rather than a long term commitment).  Ideally we'd get 347 in branch-2
>> for 2.0.4-beta and have that release to address issues that come up to
>> fix for GA.  Cloudera is actively testing 347 and parts of the
>> community are eager to pick it up so I think that would work out
>> timing wise.  Reasonable?
>
> ATM's suggestion of removing HDFS-2246 in trunk, but not branch-2, is
> a rational compromise: it allows some period for others to adapt, but
> not an indefinite one. It's not clear what you're proposing, if
> anything.

I'm not proposing anything new, Nicholas said he had some concerns
with ATM's proposal and we're discussing them.

Specifically, ATM's proposal does not allow for a single release that
contains both 347 and 2246 (he proposes removing 2246 from branch-2
when 347 is ready). I think Nicholas was saying that we should not
remove 2246 immediately for the sake of a transition period (which I
interpret to mean a release that supports both), I responded saying
that the transition that we'd have in ATM's proposal (people adjust on
trunk and there's some branch-2 release that flips over) is sufficient
given that 2246 was intended as a short term thing.  Curious if
Nicholas and others think that's reasonable.

Thanks,
Eli

>
> Nicholas/Suresh: have you had a chance to review HDFS-347, yet? -C
>
>> On Mon, Feb 25, 2013 at 12:50 PM, Tsz Wo Sze <[EMAIL PROTECTED]> wrote:
>>> I agree that HDFS-2246 is a short term solution and we should not keep it there forever.  However, we still need a transition period to replace an old mechanism by a new one.  No?
>>>
>>> Tsz-Wo
>>>
>>>
>>>
>>>
>>> ________________________________
>>>  From: Eli Collins <[EMAIL PROTECTED]>
>>> To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>; Tsz Wo Sze <[EMAIL PROTECTED]>
>>> Sent: Monday, February 25, 2013 10:24 AM
>>> Subject: Re: VOTE: HDFS-347 merge
>>>
>>> On Sat, Feb 23, 2013 at 4:23 PM, Tsz Wo Sze <[EMAIL PROTECTED]> wrote:
>>>> I still do not see a valid reason to remove HDFS-2246 immediately.  Some users may have insecure clusters and they don't want to change their configuration.
>>>
>>> Because it doesn't make sense to support multiple mechanisms for the
>>> same thing.
>>>
>>> 2246 was always intended to be a *short term solution* util 347 was
>>> completed, eg see Sanjay's first comment on 2246:   "A shortcut has
>>> been proposed where the client access the hdfs file blocks directly...
>>> This is non-invasive and is a good short term solution till HDFS-347
>>> is completed."
>>>
>>> Thanks,
>>> Eli