Re: Upgrade Plan
Date: Tue, 20 Jan 2009 21:40:41 -0600
Message-ID: <3Pwdl.11782$as4.3926_at_nlpi069.nbdc.sbc.com>
Mladen Gogala wrote:
> On Mon, 19 Jan 2009 14:13:16 -0600, Michael Austin wrote:
>
>> I came from the DEC/VMS world -
>
> Interesting. So did I. It was a long time ago, my last VMS was 5.5-2 on
> VAX 6640. I consider VMS to be the best OS ever made. I am not, however,
> sorry that DEC has failed because those #$%! have well deserved it with
> the story of DECSystem, MIPS and Alpha. I'll tell you all about it over
> a pint, should an occasion arise.
>
>> the one true cluster that no one has >> been able to duplicate in the more than 25+ years that the DEC cluster >> has been around and in conjunction with the formerly-DEC-now-Oracle Rdb >> was what Ellison always dreamed his RDBMS would become.
>
> Actually, if you have ever worked with the Tru64 UNIX clusters, you will
> find out that Oracle processes in the new CRS system correspond with Tru64
> on one-to-one basis. Essentially, I believe that HP has turned the code
> for those processes over to Oracle and that Oracle has adjusted them to
> Linux. At least that is what my private sources within the Oracle Corp.
> indicate that has happened.
One thing I did not say earlier is that, as you have noticed, CRS **IS** the DLM/cluster bits from Tru64 which got those bits from - you guessed it... OpenVMS. When HP bought Compaq/DEC, there appears to have been some back room negotiations that resulted in the death of Tru64 and Oracle getting the cluster bits. (this confirmed from an anonymous :) Oracle engineering source).
>
>> His many >> attempts at OPS failed miserably even on VMS.
>
> I would disagree completely here. I was working for Oxford Health Plans
> from 1999 - 2004, when they were bought by United Health, on a very
> successful OPS implementation. Arup Nanda was working with me, so the DBA
> team was fairly strong, but it was a successful implementation of Oracle
> OPS on Oracle 8i/ 4-way HP 9000/N.
>
>> When 9iRAC was first >> introduced back in 1999, the Oracle engineering rep at that time stated >> categorically, that it only worked as was originally designed on 2 >> platforms. OpenVMS and Tru64 UNIX clusters - because they are the ONLY >> ones that can natively mount ALL storage devices concurrently.
>
> Oracle8i OPS did not work anything but raw devices. It wasn't an easy
> setup to do, but it was possible.
>
>> ASM and >> CRS make that a bit easier, however, having a single set of OS files and >> directories across the cluster will continue to elude the UNIX market.
>
> Actually, it is already available. There are at least 2 open source
> cluster file systems (OCFS and GFS. GFS is quite interesting and Matt
> Zito of GridApp, one of the best RAC guys in the world, recommended it to
> me a long time ago), there are also commercial ones (VxFS/VCS, HP
> PolyServe) and, of course, there is a native implementation of JFS2 on
> AIX. Oracle pushes ASM because it's unusable for competition. If you put
> your database on VxFS, it can also be used by UDB or MySQL. If you put
> your files on ASM or Exadata, Oracle is certain that you're a true blue
> Oracle site and that they have a vendor lock.
>
Again, not anywhere as seemless and easy as configuring an OpenVMS cluster... :)
After witnessing the pain of trying to configure VxFS/VCS, a pure RAC/CRS/ASM environment is far more simple, attractive and less expensive. Having 2 cluster managers competing against each other is usually a disaster waiting to happen. Especially when people forget which library needs to be included/replaced :) Received on Tue Jan 20 2009 - 21:40:41 CST