A long time ago, in a galaxy far, far away, tsurkin@mail.jct.ac.il (Tsirkin Evgeny) wrote:
> Hi list!
> We are using currently postgresql 7.3.4 havily and planning an upgrade.
> our choices are either 7.4.x which comes with distribution (suse) or 8.0.1 .
> We can allow our selfs some testing time and minor changes in applications
> if needed but of course this is good to avoid .
> So the question is if the 8.0.1 is really well tested ,or maybe we
> should move
> to 7.4 or just leave 7.3 where it is?
> We wil be thankfull for any sharing of experience .
> Thanks.
I would certainly seriously consider a move to 7.4.x; that should be a
useful improvement in terms of both functionality as well as bug
fixes.
Version 8.0.1 is a bit less widely used, just yet. Two things about
it cause me a bit of pause:
1. Introduction of native Win32 support has "touched" quite a lot of
code, with some corresponding potential for diminishing the
ability to consider past testing to remain legitimate;
2. The TWO changes in cacheing are pretty major, namely the
implementation of ARC, with its forthcoming replacement with 2Q.
Tom Lane has observed that there have not been many new defects found,
and that of those found, many actually do apply to earlier versions.
Which is eminently encouraging.
For systems not expected to go into production for a while yet, I'd
personally be fairly game to use 8.0.2, once released. For things to
"go to production" immediately, I would be more than happy to use a
7.4.6 or 7.4.7; you may fairly reasonably expect to see material
improvements with a 7.3->7.4 transition.
--
(format nil "~S@~S" "cbbrowne" "gmail.com")
http://linuxdatabases.info/info/slony.html
Rules of the Evil Overlord #173. "Although it would provide amusement,
I will not confess to the hero's rival that I was the one who
committed the heinous act for which he blames the hero."
<http://www.eviloverlord.com/>
