Hi Simon,
No apologies are necessary on this list. I didn't think you sounded
anything but puzzled. I think others are too.
I perhaps sound a bit exasperated because I've tried to explain this
before. We aren't holding anything up for some magic reason. We ship new
fixes as soon as they have been fully tested. Here's a process description
which may help :
We fix problems as soon as we can, the fixes get rolled into the "build",
and when our turn comes to get a crack at those very precious resources in
System Verification (testing costs time and money and it takes many people
and machine resources to do it), we give the Verification folks a new
product images which they begin testing. They usually get several new
images over the course of the test cycle to fix anything they found broken
when they tested it, new or old. When they think the new image is OK, we
release it to Distribution in Boulder, Colorado, and they release it to the
world within 48 hours.
So an efix is from the start of the cycle and has not been more than "unit
tested" by anyone, which is why we don't just throw them out on web site
and say "come and get 'em". We have seen a lot of 'em fail when
integrated into the product. I wrote one myself which ran great on AIX and
cored immediately on Digital and Solaris.
Cheers
James Shanks
Tivoli (NetView for UNIX) L3 Support
Simon Long <simon@nettrack.com.au> on 01/27/99 05:50:32 PM
Please respond to simon@nettrack.com.au
To: James Shanks, Netview Forum <NV-L@UCSBVM.UCSB.EDU>
cc:
Subject: Re: e:fixes
Hi James,
Thanks very much for your response mate. I apologise if I sound
untrusting but working for an IBM business partner I sometimes feel
a bit on the periphery of things. I also apologise to Tivoli support if I
come
across sounding ungrateful ( I must admit it did read like that ). I'm
trying to get a feel for things so I can give decent advice to customers.
Regards,
Simon Long
Nettrack Technical Solutions Pty Ltd
James_Shanks@TIVOLI.COM wrote:
> What story are you expecting?
> An efix is a new executable which, on the developer's box at least, fixed
> the reported problem. But it has not been through formal test or system
> integration, and it is never going to be regression tested against the
> level of code you actually have -- it will be incorporated into a new
level
> and that will be tested. So of course the response is "don't apply
unless
> we tell you". Level 2 Support does not want to get a frantic call from
> someone who blindly downloaded incompatible efixes and installed them and
> now cannot get the system to run at all. They'd like to help verify that
> the problem you have is likely to be solved by the efix you want and then
> to track whether or not it really did. They want to make certain that
you
> understand that efixes are experimental, and may not work, and that you
> need to have a backup in place.
>
> James Shanks
> Tivoli (NetView for UNIX) L3 Support
>
> Simon Long <simon@nettrack.com.au> on 01/26/99 09:00:15 PM
>
> Please respond to simon@nettrack.com.au
>
> To: NV-L@UCSBVM.UCSB.EDU
> cc: (bcc: James Shanks)
> Subject: e:fixes
>
> Hi,
>
> I have to support clients running Netview on AIX, Solaris, and
> Netview platforms. I constantly hear about e:fixes and have applied
> the
> IX81622 for security on Solaris - BUT - I have yet to receive advice
> from Tivoli that doesn't say "Don't apply any e:fixes until we tell you
> to".
> This is OK but I can't help thinking this is not the full story. Can
> anyone
> help me?
>
> Thanks heaps,
>
> Simon Long
> Nettrack Techynical Solutions Pty Ltd
|