April meeting recap and next meeting sneak preview!

Hello!

Welcome to another edition of "what happened at PDXPUG". This time
with more exclamation points!

Ian's slides are at: http://znark.com/slony/

Summary of content:

* Turnout to the meeting was great - 12-13 people?! rad!
* Ian's talk on Slony
* Devin raises llamas and emus
* Next meeting (please scroll down!)

------

We got the meeting started right at 7pm with introductions. There
were a few new faces (Hi Devin, Chris and Jay!).

We didn't have an EXPLAIN operator of the month, but it sounds like
we will have one next month.

No reports were made of MySQL sucking.

--

Ian dove right into Slony and his implementation of it at RenTrak. He
used Slony to replicate a portion (700GB) of a large database
(1.5TB). Data replicated to slaves is used for reporting. He had two
slides devoted to the pros and cons - pros included that its free/
open-source, it provides a reasonable disaster-recovery strategy and
can help you upgrade a database; cons included complexity, no direct
support for schema replication, and no support for disconnected
operation. Enemies of Slony include very long running transactions
(including VACUUMS on huge tables), and tables without primary keys.

There were quite a few very useful slides concerning the setup and
configuration. Ian went into some detail on the promotion process and
what happens when a slave/master lose their connection. The ability
to specify replication sets that are subsets of a database for update
was cool. There were quite a few more configuration details - please
see the slides for them!

Ian also explained a trick for getting around exclusive locking when
applying script changes - move the affected tables into a smaller
replication set temporarily, make the change (with EXECUTE SCRIPT),
and then move back into the production replication set.

Ian mentioned that the documentation for Slony is pretty good - you
can check it out at: http://slony.info/documentation/ . He also
recommended using the mailing list and that development effort has
been quite good.

An excellent and informative talk! Thank you, Ian!

--

Afterward we retired to the Lucky Lab for beer. There was a tasty
stout on nitro, and we discussed such fascinating topics as my fear
of cockroaches, the state of higher education in Oregon and the
possibilities for the next two months of meetings. There may have
been some discussion of database/programming issues at the other end
of the table.

--

On to the sneak preview:

David Wheeler has offered to give a talk on Object-Oriented Database
Design on May 15. Tom would be our EXPLAINer of the Month.

We also had an offer from James to give a talk in JUNE on his PhD
proposal (which was accepted! Huzzah!). We would need to move our
June meeting back a week or so, and move the meeting outside of
FreeGeek (field trip!) because James is leaving for the Seattle-area
during our normal meet-week. I'm waiting to hear back from James on
details - date and location. Finally, there is at least one
postgresql "core" person who is interested in speaking to our group -
I'm trying to figure out when a good time would be apart from OSCON -
September maybe? Feedback is welcome.

Gabrielle has graciously offered (at least I think that she did!) to
lead our meeting in May. I will be going to Eugene to talk about
PostgreSQL to a new IT users group that's just started down there.

Any other meeting ideas? Cool project you'd like to talk about?
Something you'd really like to know about? Cookies you'd like to bake
for the group? Please send them my way!

--

thank you for reading all that!

The sun is out. I'm going for a bike ride.

See you soon!!

-selena