You are responding to the wrong thread. This discussion is a spin-off from Bill's original post, and the subject has changed, The spin-off started with Don Nielsen's response, which simply asked "Where might one find good instruction on how to read a dump? This is probably my poorest skill and I should be better at it." Eventually, somebody actually changed the subject line, as Jean Snow had asked after the thread had become sidetracked.
David de Jongh
I have no dog in this discussion - I am long gone from BMC - but to put this back on the rails; what Bill Blair was specifically asking for was people who have deep z/OS development and diagnostic skills.
Dump reading is just one of those skills. In Bill's case they already have shared product infrastructure (*) that diagnoses and recovers from abends in product code, regardless of the state it is running in. That's the thing that vomits up diagnostic messages and captures LOGREC and SVC dumps. All of that is completely automatic and in most cases you can figure out the root cause just by reading the diagnostic messages.
...except when you can't. That's when you need to be able to read and analyze the contents of an SVC dump, or a SAD. None of that has the slightest bit of anything to do with LE dumps, or SYSxDUMPs. To meet Bill's requirements you MUST know your way around z/OS internals sufficiently to understand the control block chains indicating the state of the machine at the time of the error. You get even more brownie points if you can do that for dumps taken out of FRR routines for work that was locked or disabled at the time of the error.
There is a limited number of people out there who have those skills and we probably already know most of them personally through having been in the industry a long time, but he's hopeful that some keen / talented person will step up and say "oh yeah, I can do that".
CC
(*) I know this because I wrote it.
Sent from my iPad
On Jul 30, 2013, at 9:49 AM, "David de Jongh" <[email protected]> wrote:
> This was "déjà vu all over again" for me. We have an in-house abend
> analysis routine driven by ...
Dump reading is just one of those skills. In Bill's case they already have shared product infrastructure (*) that diagnoses and recovers from abends in product code, regardless of the state it is running in. That's the thing that vomits up diagnostic messages and captures LOGREC and SVC dumps. All of that is completely automatic and in most cases you can figure out the root cause just by reading the diagnostic messages.
...except when you can't. That's when you need to be able to read and analyze the contents of an SVC dump, or a SAD. None of that has the slightest bit of anything to do with LE dumps, or SYSxDUMPs. To meet Bill's requirements you MUST know your way around z/OS internals sufficiently to understand the control block chains indicating the state of the machine at the time of the error. You get even more brownie points if you can do that for dumps taken out of FRR routines for work that was locked or disabled at the time of the error.
There is a limited number of people out there who have those skills and we probably already know most of them personally through having been in the industry a long time, but he's hopeful that some keen / talented person will step up and say "oh yeah, I can do that".
CC
(*) I know this because I wrote it.
Sent from my iPad
On Jul 30, 2013, at 9:49 AM, "David de Jongh" <[email protected]> wrote:
> This was "déjà vu all over again" for me. We have an in-house abend
> analysis routine driven by ...
