Patch installations

[Willem Grooters]

John,


Again, PLEASE, PLEASE report any problems you experience so they can be properly diagnosed and corrected (if found to be a fault in the product). It does nobody any good to assume the cause of a particular problem and adopt overly restrictive policies as a consequence.


There is NO way to do this IF YOU DON'T HAVE A SUPPORT CONTRACT.

As I pointed out several times before (including the ITRC BoF during last bootcamp): Contract or not, if something is rotten in a patch, ANYONE should be able to report it (and get support if the problem really is within the patch).

Willem

[John Gillings]

Wim,

>But no, it is necessary to do it because even HP makes errors

We're not perfect, but then I'm not aware of any recent patch or upgrade which broke upwards compatibility and made relinking necessary. That's not to say it's not possible (I seem to remember an RTL patch in the early V7's that caused some trouble - an isolated case), but to make relinking a policy "just in case" is rather extreme.

Willem,

>There is NO way to do this IF YOU DON'T HAVE A SUPPORT CONTRACT.

There is. At the top of this page, and every page in ITRC there is a "contact hp" link. From there click "ask a question about ITRC" on the contact page choose "Maintenance/Support (Compaq products)" and "Downloading: patches and software". Enter the description of the problem you've found.

Granted this is not as effective as logging a case with a customer support centre, but if you haven't paid for support, what do you expect?

[Volker Halle]

John,

sometimes it's not so easy to identify (application) problems induced by 'installing a patch' (especially if you installed more than one). The system manager may be more forced to concentrate on getting things running again and this may be contraproductive regarding analysis of the problem and then documenting and reporting it.

Sure, it needs to be done and reported so that it gets at least a chance to be fixed. Problems HP does not hear about won't get fixed.

Volker.

[Uwe Zessin]

Agreed. Many customers don't even have a test system. And once the system is working again there is no more time to 'play' with it.

Happened to me last weekend. We wanted to check out some things with advanced server, but it refused to start. Took some time to find out that the PWIP component of TCP/IP didn't start either. Apparently the TCP/IP ECO blew it.[*]

I have re-installed the base version and it worked again. It was a long night (we left sunday morning at 01:30 to get some sleep) and if we are lucky we can continue our work in september.

We decided to do the upgrade in two steps so we can fall back in case we hit a wall.

[*] I have later read on news:comp.os.vms that this is a problem when the name resolver is not set up - guess what isn't running here?

[Jan van den Ende]

John,


no need to re-link.


Going from 7.3-1 to 7.3-2 an applic which was linked against a host of Progress objects simply failed to start. (And yes, I know, Progress is out of VMS as per 01/01/01, but the followup applic, a tru64 one, still is not operational).
And YES, after re-linking, it did start.
However, this app is the app that services communication (using mailboxes) with remote systems to two cooperating local apps, and it DID communicate with one, but NOT with the other.
We re-linked the one not communicating, and everything is now running under 7.3-2.

And yes, we took the precaution to copy the complete app, and re-link it into a new "dash-version". A but confusing to the supporting supplier (whose specialist was on holiday at the time!), but WE have got separate versions.
And it is VERY reproducible: if we try to activate the old comm app (against the test database), then it WON'T start.
If we run the new comms app against the test database, and the old local app against its test db, then they both run, but they don't communicate with one another.

The one thing that will be very hard to demonstrate, is for them to run under 7.3-1 again., but we have tons of day-by-day tracefiles and logfiles, for the last 5 years, to tell the story of running under 6.2, 7.1-2, 7.2-1, and 7.3-1.

Very hard to reconcile that with "no need to re-link"

I WILL try and find some time for a formal report on this.

Jan