[Default] On 30 Aug 2017 07:22:46 -0700, in bit.listserv.ibm-main lionel.d...@va.gov (Dyck, Lionel B. , TRA) wrote:
>Clark - an active project to get off JES3 - that is easier said than done. >There is no cost justification for doing so, there is no ROI, there is no >significant benefit to such a project. All a site will encounter is >significant costs to (a) retrain sysprogs, operators, all JECL users, (b) >convert from using JES3 DJC to the not quite the same JES2 DJC networks, (c) >loss of data set awareness allowing jobs to enter execution only to wait for >data sets held exclusive (disp new/old) or trying to access data sets >exclusive, (d) loss of functionality from locally developed DSPs, (e) >conversion of all JES3 user exits with probable loss of functionality, (f) >updating all automation tools to use JES2 commands (assuming there are >equivalent commands), (g) the lost opportunity to use the resources spent on >the migration for other projects that help the company be more competitive, >more productive, save money, be more agile, improve operations, support their >users/customers, etc. I was at a shop that went from HASP to JES3 when we went from MVT to MVS. JES3 was supplied by headquarters (I was at a division). When the division I was at was sold to another company I inherited all of the JES3 customization. To cut the work of maintenance, we evaluated all of the mods (we were already using CA-1 instead of the corporate tape management) and eliminated those that were not applicable to our shop. A real push to go to JES2 which was the corporate standard for MVS systems at the new company came when we wanted to use SNA-NJE to talk to the VM/VSE corporate headquarters. We also were apparently the first to use Bi-sync NJE with VSE since the handshaking didn't work and the IBM Network Protocol Board had to get involved. Management decided that CA7 and 11 were needed to better control job submission and CA-Dispatch was needed to better manage our reports. These choices made the extra costs of JES3 hard to justify. Several things made the conversion relatively simple and the conversion forced a full implementation of CA7 and 11 as well as CA-Dispatch because these were needed to get rid of some of the JES3 JECL. Also helping was the fact that I was monitoring PROCLIB concatenation usage in one of the JES3 exits, something I also implemented in the JES2 exits we created. All JES2 modifications were done in exits and are in file 432 on the CBT tape. These included setting test job classes based on submitter, CPU time, TCAM queue usage and number of tape drives needed all done in a modification of an EXIT6 found on the CBT tape (I forget the company it is so long ago but it is in the comments on the CBT tape). We handled tape drive contention by automating the responses to drives not available to do a WAIT,NOHOLD. Given that JES3 costs more than JES2 and sometimes has not gotten function at the same time as JES2, it probably would be worth while for organizations to review what functions they still are using and seeing if some of them are already provided in either "base" zOS or other products already in the shop. This effort can be a precursor to a conversion or worthwhile in its own right if it eliminates organization unique code that has to be maintained. A report management system probably is something every large organization should have. Is a job scheduler enough better at controlling jobs than DJC networks that it is worth the cost? I liked JES3 but in the shop I was at, it became an unnecessary cost and a political burden. Clark Morris > >Just my $0.01 > >This e-mail reflects only my opinion and not those of my employer or those I'm >contracting with. > >-------------------------------------------------------------------------- >Lionel B. Dyck >Mainframe Systems Programmer - TRA > > >-----Original Message----- >From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On >Behalf Of Clark Morris >Sent: Wednesday, August 30, 2017 8:56 AM >To: IBM-MAIN@LISTSERV.UA.EDU >Subject: [EXTERNAL] Re: IBM to stabilize JES3 (was: IBM to finally drop JES3) > >[Default] On 30 Aug 2017 04:43:59 -0700, in bit.listserv.ibm-main >rel...@us.ibm.com (Peter Relson) wrote: > >>It is very disturbing to see someone use an inflammatory representation >>such as the initial subject of this thread when that is very much *not* >>what the SOD said. > >The handwriting was on the wall decades ago on JES3. SMS was made available >to JES2 shops on XA but not JES3. JES3 required the licensing of BDT to get >SNA NJE while JES2 had it native. SNA NJE between JES3 and VSE did not work >requiring resolution by what I recall was the Network Protocol Board (I was >the systems and JES3 programmer at the shop that was first to try it). Over >the years the more expensive job entry sub-system was the last to get some new >functions if it ever got them. I would highly recommend that JES3 shops have >an active project to get off JES3 and save money. > >Clark Morris >> >>Maybe in practice stabilization will result in some JES3 users choosing >>to move (perhaps because they need new function that would become >>available only in JES2), but IBM is not dropping JES3, nor did the >>statement of direction say or imply anything about doing so. That's >>like saying just because we might have stabilized some system service >>that you must stop using it. That too would be a faulty conclusion. >> >>Regardless, input such as what Cheryl W refers to is important. >> >>(John Eells would probably have stated the above in a cleaner way; >>apologies to him.) >> >>Peter Relson >>z/OS Core Technology Design >> >> >>---------------------------------------------------------------------- >>For IBM-MAIN subscribe / signoff / archive access instructions, send >>email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, send email to >lists...@listserv.ua.edu with the message: INFO IBM-MAIN > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
