Re: Damaged IBM Cross Reference Files

Larry implied the problem _would probably_ be reported [as in the future, with no note of what was done in the past], so why surmise "IBM knows about it"?

I do know that issues with ALIAS had remained since I last worked on any of that code. Those were best resolved by always treating ALIAS as permanent references, rather than as temporary overrides. An ALIAS is implemented as a permanent object, a DDM File object. There is no concept of REPLACE() for the database cross-reference code, yet the DDMFile features [since that is a /device file/ type], had to enable the /replace object/ feature. That the ALIAS code uses that flawed interface, increases chances of seeing errors; more likely, by actual use of the CRTDDMF REPLACE(*YES). So... If an override is required, just use OVRDBF instead of an ALIAS. Besides, creating a permanent object is expensive for something that is then just discarded, whereby deleting an object is also a relatively expensive operation; obviously that the object details are inserted and then deleted from the dbxref makes them even more expensive than any objects not similarly tracked.

FWiW there is little if any /trace/ code for the *dbxref feature; it is IMO not very serviceable. My efforts advocating trace records in any database code were for naught. And my idea to enable tracing the interface between the /client/ and the /db x-ref server/ by sending journal entries to its journal, was implemented instead as the limited server-only information in the flight recorder that I had mentioned in another post.

Regards, Chuck

Alan Campin wrote:

This whole thing puzzles me. From what I can determine, it is a
on-going problem that IBM knows about but it has never been
fixed. The why is the puzzle.

On Wed, Jul 15, 2009 at 11:24 AM, Ketzes, Larry wrote:

He is doing a lot with SQL tables. I wish I could isolate what
exactly is the cause, and proactively put some type of 'Best
Practice' in. that is assuming he is doing something SQL wise
that is not 'Best Practice'.

I'll probably open a ticket with IBM. Maybe a trace will show
something.