Thread: Re: Migrating from MaxDB to postgresql

Re: Migrating from MaxDB to postgresql

From
Christopher Browne
Date:
Quoth John <no@email>:
> As per subject, i'm considering migrating a database (still in
> development) from MaxDB to postgresql. The main reason for this is
> that the stored procedures (functions) in MaxDB are unreliable and
> hard to debug, and that the JDBC driver is still experimental.
>
> I thought I'd post here to find out how well functions and JDBC are
> supported in postgresql.
>
> If anyone has info or experience on these two things, I'd be
> interested to see what your opinion is.

Some internal "hackery" has occasionally taken place with JDBC drivers
that have caused internal consternation (basically due to developers
needing some new features that were in the "beta" JDBC code), although
that seems to be fading with 7.4.  And if you don't hack on the
drivers, you won't get bitten by that sort of thing :-).

The fact that there are easily multiple levels of quoting in stored
functions certainly makes debugging a bit of a pain, but unreliability
is not one of the problems observed.  For instance, the Slony-I
replication system <http://slony.info/> has a considerable portion of
its code that consists of pl/pgsql stored functions, and it would
break with great horribleness if stored functions were at all
"flakey."
--
"cbbrowne","@","linuxfinances.info"
http://www.ntlug.org/~cbbrowne/
Just because you're paranoid doesn't mean they AREN'T after you.

Re: Migrating from MaxDB to postgresql

From
John
Date:
Christopher Browne wrote:
> Quoth John <no@email>:
>
>>As per subject, i'm considering migrating a database (still in
>>development) from MaxDB to postgresql. The main reason for this is
>>that the stored procedures (functions) in MaxDB are unreliable and
>>hard to debug, and that the JDBC driver is still experimental.
>>
>>I thought I'd post here to find out how well functions and JDBC are
>>supported in postgresql.
>>
>>If anyone has info or experience on these two things, I'd be
>>interested to see what your opinion is.
>
>
> Some internal "hackery" has occasionally taken place with JDBC drivers
> that have caused internal consternation (basically due to developers
> needing some new features that were in the "beta" JDBC code), although
> that seems to be fading with 7.4.  And if you don't hack on the
> drivers, you won't get bitten by that sort of thing :-).
>
> The fact that there are easily multiple levels of quoting in stored
> functions certainly makes debugging a bit of a pain, but unreliability
> is not one of the problems observed.  For instance, the Slony-I
> replication system <http://slony.info/> has a considerable portion of
> its code that consists of pl/pgsql stored functions, and it would
> break with great horribleness if stored functions were at all
> "flakey."

Cheers Christopher.

I'll try and resist the hackage and get started with the JDBC. The
quality of the documentation seems good, so I'll have a go at
translating some stored functions this afternoon.

John