[Bucardo-general] Bucardo question

[Женя Деревянкин]

Thanks for the quick and full answer! I's much more then I've expected!
And thanks for noting the possible problems with Debian - It's going
to save me nerves on testing.

So, my goal is to make several "clone" databases (3 or more), which
must be kept in sync even after one of them fails. After recovery the
database can be put in operation manually.

But if the server that fails is the one bucardo is running on? What
must happen then? I mean not just that postgres dies, but the whole
server. A new instance of bucaurdo should start running on one of the
remaining servers, keeping the data synced? Won't that interfere into
the database structure?

Or maybe there's a better way of doing this using some default tools
or other replication packages?

Sorry for the dumb questions, but i'm a newbie in high availability systems :)

--
Best regards
Eugen Derevyankin

2009/12/6 Goran Gugic <goran.gugic at gmail.com>:
> The answer is yes and no. :)
>
> Replication can get pretty complex pretty quickly and there are many caveats
> - I would say that it might not even be realistic to have fully automatic
> replication (for all exceptions - database crashes, application errors and
> bugs, duplicate keys and ref integrity violations, connection errors,
> etc...)
>
> Extra shell code is strictly speaking not necessary for asynchronous
> multi-master replication as long as there are no exceptions.
>
> It might be necessary if you get into trouble (broken connections during
> replication) or you might opt to fix things semi automatically (by managing
> the bucardo manually, just to illustrate what I mean - there is a
> onetimecopy field in sync table with which you can tell bucardo to catch-up,
> or actually to replicate everything from master to the slave once things are
> fixed - this is strictly speaking no extra shell code, but require you to
> modify bucardo settings restart it, etc...)
>
> Wiki might seem confusing, however you really should read it until all of
> the concepts are clear - and preferably until all of it is clear (all of the
> pages on http://bucardo.org/wiki/Category:Bucardo, and of course don't
> forget to read install and readme)
>
> Installing and running the tests during the install would be the first step
> (the tests might not run on debian based distributions if you use standard
> postgres package due to debian organizing postgre databases into clusters)
>
> I'll cc this answer to bucardo-general so that people who know more than me
> correct me.
> Actualy, now that this is on the list, a questio: is the removal of links to
> http://bucardo.org/bucardo.html deliberate? I found that document to be
> really nice introduction to the system with no equivalent in current wiki.
>
> Good luck,
> Goran
>
> On Sun, Dec 6, 2009 at 1:10 PM, Женя Деревянкин <jekader at gmail.com> wrote:
>>
>> Hello, Goran!
>>
>> I've seen your posts on the Bucardo mailing list, and you seem to have
>> the same scenario I do - I need to set up a few equally important DB
>> servers, that should have all the data in sync.
>>
>> Could you please tell me if that can be acheived using Bucardo, since
>> their wiki is not working properly, and I can't get the whole picture
>> from there.
>>
>> If it is possible, does it work out of the box, or requires extra
>> shellcode? I need all the servers to exchange data updates, and if one
>> fails, after recovery it should be filled with data from the remaining
>> ones.
>>
>> Thanks in advance, and sorry for disturbing.
>>
>> --
>> Best regards,
>> Eugen Derevyankin