Henrik Lindberg wrote:
We asked the same question some weeks ago and got a positive answer back.
We have things set up this way now, and are in the process of fixing issues
with things that are different when running on the virtual DB as opposed to
running directly on postgres. (In our case postgres sequence numbers and
ending up in infinite loops in sequoia).
Interesting - hope you get it sorted!
When we have this fully working, at some point we will use (at least) 3
machines to make sure we do not have a single point of failure when we
perform maintenance on one of the nodes.
That's a very point - maintenance should be possible without too much
risk...
I'm interested in the performance of Sequoia in this setup, versus just
using a single 'naked' database.
In theory, reads will be quite a lot faster (up to 2x when high load),
and writes a little slower (as both controllers need to finish the write).
But my question is - how much faster, and how much slower, are these
operations turning out to be? And what is the performance hit for using
Sequoia, is it significant?
I'm about to do some tests to answer these questions for myself, to some
degree, but would love to know the experience of others...
thanks,
Alex
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
