incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Samuel Newson <rnew...@apache.org>
Subject Re: Migration from bigcouch to later couchdb
Date Sat, 01 Feb 2014 13:06:00 GMT

Sure, that’s the fallback position. CouchDB’s on-disk format has changed several times
over the years, and each time we teach the code to upgrade old to new as it encounters them.
We will do that for current CouchDB to clustered CouchDB, the question here is whether we
will do the same for BigCouch 0.4 dbs. We will certainly try, and it might be no extra work
at all, but I think it’s acceptable to say "use replication" in that case.

B.

On 31 Jan 2014, at 18:42, Florian Westreicher Bakk.techn. <stuff@meredrica.org> wrote:

> Would replication not solve that problem all by itself? 
> 
> Robert Samuel Newson <rnewson@apache.org> wrote:
>> 
>> Good question. We are certainly obliged to ensure couchdb pre-merge
>> data upgrades to couchdb post-merge but the bigcouch to couchdb path is
>> not something we’ve yet committed to.  That said, unless it turns out
>> to be intractable (which seems unlikely) we’ll certainly make an effort
>> to support that. The directory hierarchy will be the same, at least.
>> 
>> B.
>> 
>> On 31 Jan 2014, at 16:45, Vladimir Ralev <vladimir.ralev@gmail.com>
>> wrote:
>> 
>>> Hello again,
>>> 
>>> Just wondering if bigcouch data files will be compatible with future
>>> versions of couchdb now that the projects are merged. Will there be
>>> separate ports for data and shards? Will the directory structure of
>>> bigcouch be used or it is completely new? What would be the
>> recommended
>>> option to migrate data from bigcouch to new couchdb?
>>> 
>>> Thanks.
> 
> -- 
> Sent from Kaiten Mail. Please excuse my brevity.


Mime
View raw message