Re: ORACLE7 list of features

In <1993Feb26.153447.29626_at_netcom.com> jbennett_at_netcom.com (Joseph Bennett) writes:
>Okay, so you can't used stored procedures - then what good are stored
>procedures if you can't store them. You have many new features, but they
>will be executable only by compiling them at runtime. Also, you may not
>legally be able to use Oracle's packaged procedures if you do not buy it.

Funny, we are getting it for free because we bought TPO.

>>>Use of the same database on multiple machines require the Parallel Option.
 

>>Funny we do that now with a vax cluster just fine
 

>We do it on our Cluster too, and to be fair - if you use the parallel server
>now, you will get it free in the Oracle7 release, but basically, if you
>weren't a parallel user before - you have to pay for it later.

Not everyone would want this feature, should everyone pay for Oracle's development?

>>>Use of the Multi-Threaded Server requires SQL*Net v2 which is still in beta.
 

>>SQL*Net is in beta on what machines?
>>Not true on all machines, for example MVS has had multi-threaded servers
>>since v5 (to be fair, I only found this out yesterday).
 

>If you read CLOSELY - you will notice that I said SQL*Net <<V2>> which is
>still in beta. And as for multi-threading, yes it has been available via
>two-task operation, but not in the 'easy' way being touted in the manuals and
>in the advertising hype.

An Oracle DBA is not a job for the squemish, true.

>>>Lastly, as of now, at least on the VMS platform, only the base development
>>>tools are available (SQL*Forms, Menu, Report*Writer). NO end-user tools are
>>>available (at least from Oracle - such as Data Query or SQL*QMX or EASY*SQL).
>>>Also, the CASE tools are, as of yet, unsupported, and I was told that they
>>>would be available in 'maybe March' (which means 'maybe August'). Also,
>>>none of the financials or other 'applications' from Oracle are supported
>>>yet on Oracle7.
 

>>This is blatently false, I use case, Oracle makes SQL*QMX and EASU*SQL they
>>have been around for ages, and if this is your idea of an end user tool
>>then you must have some real clever end users.
 

>I would point to the OLS for what software is supported on Oracle7. You will
>find that none of the case tools are listed - or maybe you like using
>unsupported products - PS - Case*Designer DOES NOT work on Oracle7 - I tried
>and Oracle support agreed that it would not work and that a fix was on its
>way.

I'll take your word, but if true then Oracle screwed up on that one.

>>>There are MANY problems with the migration - I was told that the average time
>>>to migrate was 1.5 hours per 500mb of data. Our time to finish (and we
>>>ended up giving up and will rebuild our database later) was 20 hours for a
>>>2.5gb database.
 

>>I would question the wisdom of anyone who would give you the first quote and
>>would even go as far as to say that if you did it in 20 hours then it went
>>pretty smooth (considering the size of the database).
 

>If you think 20 hours is pretty smooth - and it wasn't smooth - the actual
>elapsed time for our failed effort was 4 days - then I think you are a very
>patient person. We are going to rebuild the database because we feel that
>it will be faster and safer than the migration. As for questioning my
>wisdom, I wonder whether you understand my point, which is - Oracle7 is
>in it's infancy, with none but the 'standard' development tools available
>and SUPPORTED. While the kernel may be well tested, it is my opinion that
>the migration instructions and the process leave a lot to be desired, and
>unless you have the luxury to test your migration on an exact duplicate of
>your database, be wary, it will take a long time.

I thought you meant 20 elapsed hours, as far as the supported part, I was told last thursday in a meeting with an Oracle rep that forms 30 and the other development tools would run with no problem.

>One last thing - our production database is 10G - how long should we have
>waited to migrate that???? 2 weeks maybe?

Given the fact that you have done it once and hopefully learned from anything that went wrong, I would say that you should be approaching the "average" but not there yet, say 2.5 hours / 500 mb. Puts you back around 20 hours, Either a weekend or a real long day.

                                Craig