Transoft Sponsor Message

June 2003

Net.digest summarizes helpful technical discussions on the HP 3000 Internet newsgroup and mailing list. Advice here is offered on a best-effort, Good Samaritan basis. Test these concepts for yourself before applying them to your HP 3000s.

Edited by John Burke

The previous month wasn’t an aberration. For the second straight month the number of off-topic postings to 3000-L, particularly rants, was way down and the technical content was way up. Among the off-topic items of interest, Alfredo Rego shared this from an article in the Financial Times: “IT workers knowledgeable in mainframe operations are a dying breed, although the hardware they run has proven surprisingly resilient to extinction… Experts say organizations are facing pressure to find adequately trained staff to maintain their mainframes...”

Then there was the report about the Multnomah County, Oregon Heath Department looking for a Klingon interpreter for patients who speak Klingon. And, of course, the report that New Hampshire’s famous rock formation “The Old Man of the Mountain” finally succumbed to the erosion that created it. And, finally, there was the report about Microsoft developing for the UK market a portable toilet with Internet access called the “iLoo”. Was it really a hoax as Microsoft later claimed?

In mid-August, Interex will be holding HP World 2003 at the Georgia World Conference Center in Atlanta, Georgia. As co-chairman of SIGMPE, I’d like to invite everyone to join us at the last SIGMPE meeting before end of sales for the HP 3000, and possibly, the last SIGMPE meeting ever. I hope that is not the case, but it is up to those with a continuing interest in MPE to keep the flame alive. Let’s try to make this the biggest SIGMPE meeting ever. One major topic of discussion will be the 2003 (and probably final) SIB and HP’s response. We are expecting HP’s formal response any day now, after which I hope to engage a discussion on 3000-L. At this SIGMPE meeting HP will report on its progress, and we will have an opportunity to lobby on behalf of our favorites. At the 2003 HP e3000 Solutions Symposium Mike Paivinen of HP strongly suggested that vCSY would continue to work on specific enhancements even after October 31, 2003 if there is obvious value to its customers.

I always like to hear from readers of net.digest and Hidden Value. Even negative comments are welcome. If you think I’m full of it or goofed, or a horse’s behind, let me know. If something from these columns helped you, let me know. If you’ve got an idea for something you think I missed, let me know. You can reach me at john@burke-consulting.com.

Is the root file the root of all knowledge?

This was the question: “I have an IMAGE database in which the item declarations have been added willy-nilly, and now I’d like to sort those declarations for easier readability without breaking my database and without doing a reload. Will this work?” The manager then went on to describe how he wanted to generate a schema with Adager, sort the item list, create a new empty database with this schema and replace the original root file with the new one.

Leonard Berkowitz, Patrick Mullen and Wyell Grunwald noted that since he already had Adager, he could just use the ITEM SHUFFLE command to achieve the desired result. However, they warned that any application that uses item numbers instead of item names would have to be fixed; Omnidex being one example, along with Netbase and Bridgeware (which uses Netbase).

The discussion then veered off to discuss whether replacing the root file like this would work. Several people, myself included, chimed in that we had done similar things successfully in the past. Why were we successful? Because all the logical information about the database is contained in the root, and there is nothing in the datasets except data, links and counter information.

Some people have suggested that physical information, such as the number of entries in a dataset (FORM SETS in QUERY), is contained in the root file making the suggested swap invalid. This is false. When you do a FORM SETS in QUERY, each dataset is opened to determine the number of entries. A useful side effect of this is that if you open the database in QUERY with write access and issue the FORM SETS command, each dataset has its accessed and modified timestamps updated. (If the database is only opened with read access then only the accessed timestamp is updated.)

See, HP is still working on fixes, even for obscure bugs

The key nowadays, though, is to get an HP engineer interested. One user reported that his efforts to remotely create a KSAM/XL file using RFA over a DSLINE were failing with FSERR 3, the required parameter was missing. After people checked around a bit, it was discovered this worked for CM KSAM files but not NM KSAM files (KSAM/XL).

The problem piqued the interest of HP’s Bill Cadier who reported, “It looks like this problem has been around a while. It took me some time to locate the service request filed for this, which is 5003-175117. It has not been fixed and I think the only reason for that is that nobody other than the original reporting customer ever encountered this.” A few days later, Cadier reported, “I was able to track down the source of the difficulty. It was just a simple oversight in the code. It was only checking for the CM KSAM file type in deciding whether to process the KSAM parameters into the forms message parameter that would be passed to RFA’FOPEN. The patch that will fix this will be MPEMXG5. I’ve no ETA on it as it hasn’t yet been prioritized on the lab’s to-do list yet.”

They shouldn’t have a parameter if they don’t want someone to fiddle with it

One HP e3000 user with time on his hands was looking through SYSGEN and came upon the following table (MISC -> SHOW):

He then noted that the pool controls the number of JSMAIN processes and wondered whether there was a situation where you would want to configure the minimum to be 2 or more and whether this was at all analogous to creating a pool of VT server processes.

In separate postings, Steve Macsisak and Stan Sieler provide an explanation. Macsisak said, “The main advantage is so two people can log on at the same time. If you only have one JSMAIN one will have to be created for you before you can logon. If you have a small pool then you don’t have to wait for one to be created. But having a pool of more than a few is not of much use since the OS will still keep that many in the pool all the time. Its not like the VT servers where you use the pre-built servers first and the OS will not create a new one until all of the pool has been exhausted. So for VTs you should have a few more than you will ever need which cuts down on process create overhead. For JSMAINS four or five on a very busy system will probably be enough to keep anybody from having to wait for one to be created.”

Sieler noted, “As Steve said, it affects the performance during logons. It only comes into play during two or more (nearly) simultaneous logon attempts. Thus, if your average time between logons is several minutes, you generally won’t benefit by having more than 1 spare JSMAIN. Note that you can configure the JSMAIN pool for SESSIONS independently from JOBS. If you have a lot of short jobs (e.g., averaging five or more jobs starting per minute), you might want to increase the JOBS JSMAIN pool. The same reasoning applies to sessions. In short, most people do not need more than one or two.”

HP’s Jazz Web page has a document originally prepared by HP’s Scott McClellan, but I believe updated by others, that documents operating system limits. Anyone dealing with administering an MPE system should have it bookmarked: jazz.external.hp.com/papers/limits/intro.html.

Now when did they add that?

A new user writes, “I have to write a program (in Java or C++) to externally reset a password for a user on the HP 3000. Can anyone guide me on this?” Lars Appel responded “Why not simply let the user change it with the PASSWORD command? Type HELP PASSWORD ALL at the MPE prompt to see details.” Several of us supposed “experts” were embarrassed to admit we didn’t know about the PASSWORD command and asked Lars when it was added. He responded that he was able to find it in his personal documentation system.

It turns out the PASSWORD command was documented in the Communicator for MPE/XL 3.0, which explains (though it does not excuse) why some of us missed it. I know that in my case, I went from the Classic systems to MPE/XL 4.0 and probably never read the 3.0 Communicator. After Lars’ posting, I went right to my handy ORBiT Software Pocket MPE/iX Guide and read through the list of MPE commands to see if there was anything else I was missing.

Of course it does not help that the 3.0 Communicator is not available publicly online. In fact, little prior to MPE/iX 5.5 is available online. One of the items finishing in the top 10 on the latest 2003 Systems Improvement Ballot was about making all HP documentation publicly available, even after 2006.

Another “They haven’t turned off all the lights yet” development

James Hofmeister reported from Mark Bixby’s Solutions Symposium talk “Is Your Homestead Secure?” (all about keeping your e3000 safe from hackers until 2006 or beyond), that Bixby referenced a future fix to FTP: “A future version of FTPSRVR will likely add a new parameter to SETPARMS.ARPA.SYS to globally enable or disable ‘site stream’.” Well, the future is now. The SR is 8606299646 and the enhancement is documented in FTPDOC.ARPA.SYS, which is included with the patch. All versions have limited distribution: look for patch FTPHD07 for MPE/iX 6.5; FTPHD08 for MPE/iX 7.0; FTPHD09 for MPE/iX 7.5.

John Burke is the editor of the NewsWire’s HiddenValue and net.digest columns and has more than 20 years’ experience managing HP 3000s.