-----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Ted MacNEIL Sent: Wednesday, February 18, 2009 5:29 PM To: [email protected] Subject: Re: EXECIO
My test was: 1. Two users open the same PDS (note: no E). 2. Select X; Select Y (one for each user) 3. Type something. 4. One save; two save. Whomever saved second 'WON'. I have not tried since z/OS 1.4, two years ago. So, what ENQ's do I need? We tried on the same system, and on separate systems. This PDS 'corruption' has been around longer than I have. <SNIP> Ok, being a pilot and all, I've made a landing for fuel, so I've topped off my tanks (to comply with FAA regs concerning reserves and all that). Let's back up to the mid-'90s. I have run the test Ted describes across systems and in the same DOMAIN (hey it was an Amdahl shop). The test called for ACS/WYLBUR and TSO/ISPF using the documented ISPF ENQUEUES. This was a STANDARD TEST when I was an ACS/WYLBUR developer (MVS/SP4 through OS/390 V1R1) in the mid-'90s. TODAY, we run the same situation where I work, only we are all TSO/ISPF users. We have been running this way with z/OS 1.7 through and including z/OS 1.10. Prior to working here, I have been at customer sites running OS/390 V2R9 through z/OS 1.4. One customer site was running z/OS 1.1 with three LPARs all sharing DASD. We did not have MIM. We did not get crushed PDS members EXECPT for when I made the mistake of running multiple assemblies with LNKEDTs to the same load library. There we wound up with a screwed up directory and bad members. MY OWN FAULT for NOT using DISP=OLD. Today (actually, it was during summer 2008) I ran tests with some code that I wrote using the ISPF ENQUEUES. The only time we have had a problem is when simultaneous update of the DIRECTORY is attempted (and I did my best to make sure that two STOWs happening together was part of the testing). As someone else pointed out, you will get an ABEND for that one -- system protecting you from yourself. In fact, my attempting the same old thing of multiple assemblies/link jobs now will ABEND when one BINDER is attempting to write to the same load library as another BINDER at the same time (I knew about that protection and attempted to make use of it). Also, I can, between three TSO/ISPF sessions, edit a PDS, and get the LATEST copy of a member, even though the DIRECTORY hasn't been refreshed (from my perspective). By that I mean, I can list a PDS with one account (A) and edit the same PDS from (B) and once (B) saves and ends with the xyz member, (A) can now edit xyz w/o a manual refresh. I know that this works with 1.8 and up (because those are the levels I'm logged on to). Now, there may be something happening under the covers here, because we do have TRX (only for TSO sessions). But I thought that it only cached the BLDL info for load, CLIST/REXX and panel libraries. So, Ted, as a few others have pointed out, you must have something wrong with the configuration where you are. Regards, Steve Thompson -- Opinions expressed by poster may not be those of poster's employer. -- ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
