For every release of Systems Programming for Dummies, the first rule on the first page is always the most recent change needs to disappear to correct a problem.
Since the change cannot disappear of upgrading to Yosemite 10.10.3 for various crucial tactical reasons, the only other choice is to find out what crashes it. The problem is that an intermittent crash is extremely difficult to track back to cause, particularly when it's not leaving any traces. If at all possible, OS X will write a diagnostic report for any crash and it will be accessible by launching Console. If you're not familiar with it then please get that way as that's where Apple puts crash information.
In my circumstance, the system crashes so hard there is no diagnostic information. This leaves me with nothing to review and Apple nothing to trace for similar problems, etc.
From that situation, there's nothing for it but a fundamental review to eliminate every possible variable. Two thousand bucks that I don't have are riding on this so it's eyeball-to-eyeball combat that's been going frequently 16-18 hours a day.
My purpose in writing is that it will take some while yet to fix. Bringing in all the bits step-wise is going to take some serious time and I'm estimating two-three weeks to complete it. Following the steps precisely is paramount or we will be hunting snipe forever. The trouble with hunting snipe is no-one is sure if they even exist.
(Ed: are you seriously referring to "Spin and Marty?")
Yep. Not ashamed either.
Again, not whining. Only fact. In fact, my life sucks like a mustard banana but there's no particular reason to talk about anything that sucks and ... tell me you're not doing this too ... I'm wondering if a mustard banana would work. I know jalapeño, Greek honey, and an olive make an exceptionally intricate taste so then I wonder ... mustard ... banana ... does this go?
My friend and colleague, Ronald (known nefariously under other Internet names), has worked on IBM systems for years and this is nostalgia but, wtf, sometimes that happens. When it comes to blowed-up systems, IBM has got more tools than Tim Allen on "Tool Time" and they could do just about anything anything needed to extract information from a broken system. If anyone could get blood from a turnip, IBM is the one.
Gene Amdahl was the Great God of Independent Mainframes and he had a function that permitted stopping the entire mainframe and putting it into single-step mode permitting observation much like using TEST under TSO. That guy was brilliant beyond all brilliantly brilliant.
ASTOP 80
So, your first question, Sysfrog of the Gods, is whether location 80 is hex or decimal as the machine will stop there, will not push anything through the pipeline, and will wait for the touch of your saintly hand. Frankly, I don't recall as this was thirty to thirty-five years ago. So, do you feel lucky? Do you, Sysfrog of the Gods?
We were gods. It was very good. Ronald is a bit younger so he might have missed this. I wasn't a god at that point but rather a neophyte, a godling. But it was still very good.
To be a Real Sysfrog, it is paramount that you can read system dumps. Sometimes it was said that one 'shoots a dump.' Sysfrogs love military metaphors and they're often killing something. Draw whatever sociology you like from that. Lock and load, baby (heard so many times I could not count).
And the all-time Number One of the Sysfrog Hit Parade of Throwaway Lines.
God how I love the smell of napalm in the morning.
Must be able to read dumps.
Yah, yah. Now you just call the NSA and they tell you where it crashed as they saw the whole thing. They probably caused it ... bastards.
(Ed: are you getting fooking wistful for a computer?)
In this regard, yes. When an IBM system is in trouble, you can drill so far into it that it would scare every molecule in the machine. The act is scary for you as well because you go into a world where everything is doors, windows and mirrors ... and chainsaws. MVS means Multiple Virtual Storage and this is the fundamental concept to every computing device now. In effect it means running many applications at once. Today you take that for granted even on an iPhone but this was an enormous revolution in data processing when it first made its debut.
A dump is when the operating system, every live application, every interactive session, etc is 'dumped' to disk or tape. There was a time when such a dump could be printed but that was way back and there is so much data in one now that no-one would think of printing one. Various analytical tools exist but nothing will match poring over a dump on the table in the conference room in the middle of the night. All the sysfrogs would be offering help or commiseration as it progressed. And everyone smoked a LOT of cigarettes. There was so much smoke that pall of it would hang from the ceiling as the air conditioning couldn't pump it out fast enough.
(Ed: this is nostalgia like Lawrence Welk)
Hey, no fookin' bubbles.
Since the change cannot disappear of upgrading to Yosemite 10.10.3 for various crucial tactical reasons, the only other choice is to find out what crashes it. The problem is that an intermittent crash is extremely difficult to track back to cause, particularly when it's not leaving any traces. If at all possible, OS X will write a diagnostic report for any crash and it will be accessible by launching Console. If you're not familiar with it then please get that way as that's where Apple puts crash information.
In my circumstance, the system crashes so hard there is no diagnostic information. This leaves me with nothing to review and Apple nothing to trace for similar problems, etc.
From that situation, there's nothing for it but a fundamental review to eliminate every possible variable. Two thousand bucks that I don't have are riding on this so it's eyeball-to-eyeball combat that's been going frequently 16-18 hours a day.
My purpose in writing is that it will take some while yet to fix. Bringing in all the bits step-wise is going to take some serious time and I'm estimating two-three weeks to complete it. Following the steps precisely is paramount or we will be hunting snipe forever. The trouble with hunting snipe is no-one is sure if they even exist.
(Ed: are you seriously referring to "Spin and Marty?")
Yep. Not ashamed either.
Again, not whining. Only fact. In fact, my life sucks like a mustard banana but there's no particular reason to talk about anything that sucks and ... tell me you're not doing this too ... I'm wondering if a mustard banana would work. I know jalapeño, Greek honey, and an olive make an exceptionally intricate taste so then I wonder ... mustard ... banana ... does this go?
My friend and colleague, Ronald (known nefariously under other Internet names), has worked on IBM systems for years and this is nostalgia but, wtf, sometimes that happens. When it comes to blowed-up systems, IBM has got more tools than Tim Allen on "Tool Time" and they could do just about anything anything needed to extract information from a broken system. If anyone could get blood from a turnip, IBM is the one.
Gene Amdahl was the Great God of Independent Mainframes and he had a function that permitted stopping the entire mainframe and putting it into single-step mode permitting observation much like using TEST under TSO. That guy was brilliant beyond all brilliantly brilliant.
ASTOP 80
So, your first question, Sysfrog of the Gods, is whether location 80 is hex or decimal as the machine will stop there, will not push anything through the pipeline, and will wait for the touch of your saintly hand. Frankly, I don't recall as this was thirty to thirty-five years ago. So, do you feel lucky? Do you, Sysfrog of the Gods?
We were gods. It was very good. Ronald is a bit younger so he might have missed this. I wasn't a god at that point but rather a neophyte, a godling. But it was still very good.
To be a Real Sysfrog, it is paramount that you can read system dumps. Sometimes it was said that one 'shoots a dump.' Sysfrogs love military metaphors and they're often killing something. Draw whatever sociology you like from that. Lock and load, baby (heard so many times I could not count).
And the all-time Number One of the Sysfrog Hit Parade of Throwaway Lines.
God how I love the smell of napalm in the morning.
Must be able to read dumps.
Yah, yah. Now you just call the NSA and they tell you where it crashed as they saw the whole thing. They probably caused it ... bastards.
(Ed: are you getting fooking wistful for a computer?)
In this regard, yes. When an IBM system is in trouble, you can drill so far into it that it would scare every molecule in the machine. The act is scary for you as well because you go into a world where everything is doors, windows and mirrors ... and chainsaws. MVS means Multiple Virtual Storage and this is the fundamental concept to every computing device now. In effect it means running many applications at once. Today you take that for granted even on an iPhone but this was an enormous revolution in data processing when it first made its debut.
A dump is when the operating system, every live application, every interactive session, etc is 'dumped' to disk or tape. There was a time when such a dump could be printed but that was way back and there is so much data in one now that no-one would think of printing one. Various analytical tools exist but nothing will match poring over a dump on the table in the conference room in the middle of the night. All the sysfrogs would be offering help or commiseration as it progressed. And everyone smoked a LOT of cigarettes. There was so much smoke that pall of it would hang from the ceiling as the air conditioning couldn't pump it out fast enough.
(Ed: this is nostalgia like Lawrence Welk)
Hey, no fookin' bubbles.
No comments:
Post a Comment