[Gllug] Oracle on Linux

Nix nix at esperi.org.uk
Fri Sep 17 10:51:04 UTC 2004


On Thu, 16 Sep 2004, Ian Northeast stated:
> Tethys wrote:
>> Nix writes:
>>
>>>`Oracle' and `grotesque memory hog' definitely go together. :)
>> You forgot the terms "bug ridden" and "unsuitable for enterprise use".
> 
> To be fair the RDBMS itself isn't so bad. Memory is pretty cheap now
> after all.

And it does scale quite well (not as well as DB2, and its query
optimizers *suck*, but still...)

> Their other software is somewhere on a par with the installer or
> worse. I have spent many frustrating hours with iAS, which I
> wholeheartedly do not recommend.

We call it `Internet Agony Server' because it distributes agony and
malfunctions to anyone accessing it.

(Their handling of forward and backward source and binary compatibility
in .class / .jar files is amusing --- none whatsoever. Sun helped here
by not defining a symbol versioning scheme for .class files; ridiculous,
given that it was them who came up with ELF symbol versioning!)

>                                  At least you only have to run the
> installer once per installation. You have to live with iAS until you
> can find someone else to offload it onto. And they still come back to
> bug you when they can't fix it.

And if you have a really serious bug it takes months for Oracle to fix
it (if at all) and then all they give you is an ugly workaround.

> One effect of the software being so bad is that their support staff
> (L2 and above anyway) are actually quite good IME.

Yes. My problems have been when the developers ignore what both I and
Oracle's support staff are telling them about the causes of obscure bugs
and go vanishing off down evaporatingly stupid sidetracks failing to
reproduce the bug, ignoring anyone who tells them they're up a gum tree
(not that mere mortals are permitted to speak to them at all, oh no).

(Record delay caused by that sort of idiocy: 1.5 *years*.)


But 9iAS is nice compared to Pro*C and their other front-end developer
tools. They do *everything* wrong (e.g., Pro*C does preprocessing
itself, and either loses the GCC tooldir, leading to e.g. no definition
for NULL, or finds the GCC tooldir and gets confused because the headers
in there are full of GNU C-isms. IMHO, whoever thought of making
something other than a C compiler carry out phases 1--4 of translation
(`preprocessing') should have been incarcerated before they touched the
keyboard to implement their bright idea.

-- 
`The copyright file is for everyone.  That we make it available in
 plain-text, uncompressed form rather than in spinning, throbbing
 OpenGL-rendered 3D text over a thumping dance music soundtrack is a
 feature, not a bug.' --- Branden Robinson

-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug




More information about the GLLUG mailing list