[OSM-dev] Potlatch - Richard... I know
80n
80n80n at gmail.com
Tue Mar 31 17:22:40 BST 2009
On Tue, Mar 31, 2009 at 4:26 PM, Matt Amos <zerebubuth at gmail.com> wrote:
> On Tue, Mar 31, 2009 at 4:11 PM, 80n <80n80n at gmail.com> wrote:
> > Let me see if I understand correctly what you are saying. You think that
> > the currently proposed migration from the old server and old schema to
> the
> > new server with a new schema that includes some referential integrity
> > constraints will take longer than the time that has been planned for it?
>
> i think that is what stefan is saying, yes.
>
> > I assume that Matt and co. have done some benchmark tests on the
> migration
> > and have some idea of how long they think it will take.
> >
> > Matt how long do your benchmark tests suggest the migration will take?
> How
> > much confidence do you have in your test figures?
>
> on the computer i have here (2x quad core opteron, 16Gb ram, 4x1Tb
> SATA raid0) the import takes 15h, changeset synthesis takes another
> 15h and the integrity checks + FK take about 10h, so about 40h in
> total. of course, if something goes wrong it doesn't give us much
> chance to fix it.
>
> on the new db server we're 3.5h into the import with only 1 table
> remaining, so i'm very hopeful that will finish well before 15h.
> changeset synthesis and integrity checks should get a similar speedup
> from the faster hardware.
>
> i am confident we can do the migration in the 4-day window that we're
> committed to, but i can't guarantee anything.
>
Matt
Thanks for this info. I'm reassured.
I assume that Stefan worries are based on some different scenario for
migration.
80n
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20090331/9acdffea/attachment.html>
More information about the dev
mailing list