Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Orbitz blames Oracle for site outage

Re: Orbitz blames Oracle for site outage

From: Jay <carno8rules_at_hotmail.com>
Date: 28 Jul 2003 20:23:49 -0700
Message-ID: <ab05de6b.0307281923.688deb47@posting.google.com>


mikerault_at_earthlink.net (Mike Ault) wrote in message news:<37fab3ab.0307280426.57af69a5_at_posting.google.com>...
> "Morfeo Quesoverde" <oratune_at_msn.com> wrote in message news:<3f16c994_at_shknews01>...
> > Orbitz
> > "Keith" <nospam_at_nospam.com> wrote in message
> > news:vhce9p127qe3c1_at_news.supernews.com...
> > Attached is the site message saying it is down due to the Oracle problem
> > mentioned in CNet. Anyone have detailed info on this?
> >
> >
> >
> >
> >
> >
> >
> >
> > I'm wondering WHERE, in this message, it says anything about Oracle being
> > the cause of this outage ... I can't see it anywhere.
> >
> >
> > David Fitzjarrell
>
> The most current data points to NIC problems...probably using BCM5700
> drivers...they have a rep for causing hangs in combination with other
> components on the motherborads...all sorts of posts about it on the
> DELL site and other locations...but I'll bet we don't see a retraction
> from Orbitz...
>
> Mike

  How would moving off of RAC solve a NIC issue? Did it have to do with the private interconnect? My experience with Oracle (excluding RAC) is that it is rock solid. With RAC on Sun there are at least three different vendors providing critical pieces to make it work (Sun, Veritas, Oracle). This causes confusion and finger pointing. The underlying components including cvm, dlm, and the cluster frame work along with RAC are far from an easily supportable environment. When problems happen it is difficult to get to the support people that have the in depth knowledge to help solve the problem. Oracle could probably make RAC a better product if they did support the whole package say on a Linux platform, but I don't know how many people would go for that.
  Another problem with RAC is that the database is a single point of failure. So, to get around that you have to have N*2 nodes for redundancy.
  It's my feeling that Oracle in a fail-over cluster or no cluster at all with data propagation is a much more supportable, and reliable environment. I wonder if that is the what the support staff from Orbitz was thinking? Received on Mon Jul 28 2003 - 22:23:49 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US