Re: Replication

Hi Dennis!

Some people have definitely used the existing replication implementation in
production; I hope some of them will speak up with their experiences.

I don't think much changed in the replication implementation between
2.02.006 and 2.3.

I may have removed it from the Git head at some point because we always
considered it a proof-of-concept, kind of quick-and-dirty -- it's
functional, but doesn't have automatic recovery built-in. If the connection
goes down, the replica should continue to process queries but will stop
processing transactions until you restart it as a master. You'll have to
determine yourself whether the master actually crashed or it's just a
network issue, and restart the replica appropriately to either re-establish
its connection with the master or to become a master itself.

But again, there are people happily using it in production, so I may have
been too harsh on it. :)

Cheers,
Justin


On Mon, Aug 30, 2010 at 9:32 AM, Dennis Rieks <[email protected]>wrote:

> Hi,
>
> we are using prevayler 2.02.006 for 5 years now without problems.
>
> An new requirement is replication and automatic fail-over to an backup
> server if the primary one gets unreachable. Could you tell us your
> opinion for the best solution?
>
> 1) Upgrade to Version 2.3
> There is an unfinished replication implementation. It seems to be
> removed in the latest git version. Can you tell me the reason why? Do
> you know anybody using this code?
>
> 2) A own implementation witch sends every incoming transaction to an
> queue. A background threads sends all elements from the queue to
> another server instance.
>
> 3) A filesystem approach sending new journal entrys to the other server.
>
> We need an async approach because there are a lot of transactions
> which need to be as fast as possible processed. It is no problem if
> some few transactions get lost if the primary server gets unreachable.
>
> Also there is no need for automatic resync from backup to the master
> server.
>
> For approach 2) and 3) it is not needed to update to version 2.3. It
> is practical to upgrade our code? We have no problems and don't need
> any new feature expect replication.
>
> I think 2) is the easiest way to implement replication for us. If you
> have any ideas or suggestions please let me know.
>
> Thank you
> Dennis Rieks
>
>
> ------------------------------------------------------------------------------
> This SF.net Dev2Dev email is sponsored by:
>
> Show off your parallel programming skills.
> Enter the Intel(R) Threading Challenge 2010.
> http://p.sf.net/sfu/intel-thread-sfd
> _______________________________________________
> To unsubscribe go to the end of this page:
> http://lists.sourceforge.net/lists/listinfo/prevayler-discussion
> _______________________________________________
> "Databases in Memoriam" -- http://www.prevayler.org
>

------------------------------------------------------------------------------
This SF.net Dev2Dev email is sponsored by:

Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
http://p.sf.net/sfu/intel-thread-sfd

_______________________________________________
To unsubscribe go to the end of this page: http://lists.sourceforge.net/lists/listinfo/prevayler-discussion
_______________________________________________
"Databases in Memoriam" -- http://www.prevayler.org