Sign In

Mainframe User Shop Talk The place to share user experiences, inquiries for advanced techies
z & A > Mainframe User Shop Talk > THINKIng: Do-overs Are Rarely Required View modes: 
User avatar
Dan - 9/10/2014 6:20:30 PM
THINKIng: Do-overs Are Rarely Required
DO-OVER NATION (NOT! not in System z)

Experienced mainframers solve problems using previously captured diagnostic data, which is preplanned. Other platform people and inexperienced mainframers immediately jump into performing problem recreation, do-overs and using temporary scaffolding to resolve their problem.

Which side are YOU on? How about the colleagues in your mainframe department? How about colleagues in adjacent non-System z departments?

Do you FULLY take advantage of zOS and the System z platform's 50 years of problem-solving engineering and continuous improvement and re-engineering?

If the new systems people in your organization ARE part of the do-over nation or do-over world, then that reduces the need for System z and the platform's advantages. Besides, systems availability will suffer, MTTR will increase, and user satisfaction and management availability bonuses will suffer. Right?!

Can you do a good deed, and teach a do-over enthusiast about solving problems on their first occurrence? If so, do it now. Evangelize System z. Evangelize the mainframe!

User avatar
SteveW - 9/11/2014 5:21:14 PM
RE:THINKIng: Do-overs Are Rarely Required
Right ON Dan!

If you can't fix your product issues with the gathered information just after a customer issue, then you are missing tools that we call "First Fault Problem Solving" (FFPS).

The information needed may be in trace form or event logs from the product of interest.

This is not limited to just products as System Z. We also see this information captured on some storage products as well. I am from an old company, StorageTek, and had these tools in both their disk and tape products. I know these tools are still available on Oracle/STK products as well and can say they are still working very well for customer issue resolutions.

I am "Evangelizing" the FFPS tools that could be available to all products, and I know that is what you really would like to see. Yes IBM does get this, but other companies also give these FFPS tools a lot of due diligence and importance as well.

Somehow, these tools need to be taught or shared to people that have never seen or used them.

I will try to share our FFPS tool's processes and components and ideas with anyone that maybe interested, here on this thread. I am subscribed to your POST, so I well see what requests may happen. But if people just continue to need to duplicate customer issues in order to fix their issues, I suppose they may not be interested in what you and now me are Evangelizing.

I am currently able to fix ~80% of all customer issues with a single drive trace dump taken by either our customer or CSE just after customer sees failure on our Oracle/STK tape drive. We continue to improve the trace facility when we can't fix on first fault. I can also explain how we do that as well,

I'll be monitoring this POST for questions, if anyone has any.