On Fri, 4 May 2012 03:57:24 -0500, Brian Westerman <[email protected]> wrote:
>I realize that CPUs are a lot faster than they have ever been, but using the >resources on frivolous things like keeping people from having to look up a >message by always printing the verbose text is silly. Most people have the >manuals available to them electronically, and the ones who don't probably >wouldn't be the one who is going to have to look up the message in the first >place. This is sort of funny, because there have been so many rants on ibm-main about lack of good messages in z/OS unix (and some other areas) and here IBM is trying to take one of the strong points of our platform and make it better (optionally) and there are complaints by some. I realize that it isn't IEC* messages that get those rants though. I don't have access to quickref, and again, it isn't myself I am advocating for. It is the end users / operations staff. If the messages are right there in the joblog / syslog, no one has to look up a message anywhere. And right now I am at a shop with probably about 80-90% of the entire mainframe operations staff with 5 years or less of experience. Anyway, Brian brings up a slightly interesting point to me. Can the overhead even be measured on a "normal" system. I suppose if you generated a test case with thousands of B37s in a short period of time you could measure the difference. Did IBM do this in a lab? I still don't get the concern about some extra data being written to the syslog. For the person(s) who were concerned: Can you please tell me how many lines you used to write to syslog on avg. each day and then what the difference was when you turned on the verbose messages with 1.13? I don't have 1.13 in production yet, but I just looked at the syslog from a typical / medium sized production LPAR at my client. The syslog cut at midnight last night (24 hours) has about 450,000 lines in it. I think exactly 35 messages would have generated the verbose description. So a whopping ~ 200 extra lines of syslog would have been written - an increase of .04% to store that data wherever you store it. I did the same thing for a development LPAR since I expected to see more x37 abends. I found about 100 messages there out of about 350,000 or ~ .17%. But since most are production LPARs (and things like SAP and Websphere LPARs don't even run batch), overall for the environment I would expect the average increase to be more along the lines of .04%. Regards, Mark -- Mark Zelden - Zelden Consulting Services - z/OS, OS/390 and MVS mailto:[email protected] Mark's MVS Utilities: http://www.mzelden.com/mvsutil.html Systems Programming expert at http://expertanswercenter.techtarget.com/ ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
