On Jan 8, 2008, at 11:05 AM, Chris Hoelscher wrote:

good morning everyone.
after a large software maintence upgrade, i will hoping to selectively copy afffected load modules to the runtime loadlibs using IEBCOPY - this worked well enough, but the associated aliases were not copied - resulting in what appears to be orphaned aliases in the "to" loadlibs - is there any way to tell IEBCOPY (or any other utility) to update all associated aliases in the
"to" dataset when the "root" load module is copied? or must I manually
determine what aliases must be copied along with the "root" load modules?
or am i just better off to copy the entire load library?

thanks,
-SNIP------------

Chris,

I think it somewhat depends on how sure you are that you have got all the alias(s) and a little bit of how good the sysprog is.

I always applied fixes to the maintenance pack. I then looked very closely at the linkedit output and IEBCOPY and created the s m= cards with the all the alias and primary names. I always checked twice or three times depending on what module(s) were hit. I was extra careful when the module was in LPALIB as I knew if I screwed up it would cost an IPL. If it was in a linklist library I was careful but not overly so as it might cost an abend but not an IPL (most of the time). I *ALWAYS* made a copy of the library before the copy operation. Also if it was in a linklist library I could always copy into sys1.linklib in an emergency and do a refresh. At all time I was extremely careful and if there was too many members I would set up an alternate respack also I would schedule the IPL so I was on site. If I couldn't be I would set up an alternate respack. I was really careful and double and triple checked items yea and sometimes quadruple. I was teaching a jr. sysprog and one time I let him do it without my intervention and stopped it just before he hit enter. We went over what he had done and I showed him that he had dropped an alias and he was quite embarrassed. I let him try a few more times and he got it right but I insisted that he accompany me in at a god awful hour for an IPL. I made sure he understood what could go wrong if he screwed up. He was pretty good maybe not as attention oriented as I would have liked him to be but he never screwed up and I let him fly solo a few times. One major PTF came out for JES2 and so I let him fly solo, I didn't even check on his work (well I did but he did not know it) and found a minor issue that he forgot to copy over a haspsrc member. It was not a show stopper in that it did not cause any IPL outage but it got the libraries out of sync. The PTF went on the maintenance pack with no issue and I stopped him as I did not want to have anything touch the live system pack until we were both present (I am a little paranoid). We came in at 0400 and proceeded to copy the hasjes20 module and then he said "OK lets go for the IPL" I told him NO and he looked at me and I said before you do anything lets desk check. So we went back to the desk and he looked over the SMPe output and said everything is fine and I said no it isn't and he looked at me and said ok lets check again. He did not catch the update to haspsrc I had to point it out to him. Turns out he had never run into IEBUPDTE before. To get this exercise overwith I had him copy the member in haspsrc over to the respack and I said OK. We IPL'd without an issue. I gave him the utilities manual and explained IEBUPDTE and told him that (at that time) JES2 background and he got the proverbial light bulb above his head.

The point is in my first two sentences I said to be VERY careful and only trust special people. Beginners can learn (IMO) but only by experienced people. I only trusted myself. Be very careful on who you trust. Also ALIAS(s) can kill you if you aren't careful.

Ed
----------------------------------------------------------------------
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

Reply via email to