Re: Orbitz blames Oracle for site outage

Keith <nospam_at_nospam.com> wrote in message

> So much for the 'Unbreakable' strategy..

Duh?

Re-read your sentence. *Strategy*. That implies not a product that cannot break, but an APPROACH that can deliver an unbreakable system implementation.

> Last I heard, PostGreSQL is still serving up domain names for the .ORG
> domain name registry.

Last I heard, assholes are still being born. They just grow bigger as they get older.

> Orbitz blames Oracle for site outage:
> http://news.com.com/2100-1019_3-1026450.html?tag=fd_top

And automatically you assume it is an Oracle issue?

After an o/s upgrade to a Data Warehouse some years ago, we had all kinds of problems. Db just hanging and crashing.

I spend some time with Oracle Support tracing the problem. It was a big issue. Mission critical corporate system down. It was escalated to the Munich Oracle Compotence Centre. The problem was traced... and turned out to have been caused by some friggen dumb jerk that forgot to install a kernel patch during the o/s upgrade. Async i/o was used by Oracle. And accepted by the kernel. But the kernel "forgot" to inform Oracle when the async i/o completed.

Your posting kinda of remind me of that.. especially that friggen dumb jerk part.

And this is btw not the exception.. I have many war stories about database downtime and crashes and errors. The vast majority of times it is caused by
a) operating system bugs
b) idiotic implementation that breaks not only Oracle strategies and rules, but common sense too

When the shit hits the fan.. I prefer Oracle Support. And not Microsoft Support or OEM hardware vendor support (like IBM, ICL, Siemens, etc). The Oracle support guys I have had the pleassure to work with (locally and Europe), go beyond the extra mile.

And this opinion is from bitter experience dealing with other support centres.

--
Billy