Lund Sponsor Message

April 2003

Net.digest summarizes helpful technical discussions on the comp.sys.hp.mpe Internet newsgroup and 3000-L 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

I feel like I’m copying this from last month. Actually, I am copying this from last month: “The bad news is that 3000-L was almost hijacked by several threads on politics, specifically the potential war with Iraq. So much time spent and, I’ll wager, not a single mind changed. The good news is that even though it took a little digging to find, there was still a lot of technical content to report on.” During March, however, I took it upon myself to chastise members of the list for going overboard. I think I used phrases such as “endless drivel.” Anyway, enough of this drivel; let’s get to the content.

As always, I would 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. You can reach me at john@burke-consulting.com.

They haven’t turned off all the lights yet

The following was asked on 3000-L: “I see that newer versions of Perl and Samba are available, yet I’ve heard nothing about whether or not they’re in the works for MPE. Since MPE is still supported by HP, these version updates should be coming from HP. Yes?”

Mark Bixby noted that he is working on the latest developer snapshot of Perl. As for Samba, Mark reported, “I can speak for HP here and not just myself. I have 2.2.8 already packaged in the Bixby-standard binary distribution format. All that remains before I can put it into an official HP patch is to write the ‘README.MPE’ file and conduct support training. Anybody dying to get 2.2.8 on a ‘Mark Bixby support only’ basis can contact me via private e-mail and I will make it available to you.

“By the way, the recent buffer overflow security issue is just a denial of service annoyance and does not allow execution of arbitrary hacker code on MPE. Many, many other bugs have been fixed since 2.0.7, and in combination with the new-to-MPE features of encrypted password support and improved printer support, I recommend that all Samba/iX users consider updating once the official patches become available for 6.5/7.0/7.5.”

Maybe that should have been version 0.9999?

Last month I noted that version 1.0 of QCTerm had finally been released. Wirt Atmar reported, “If you downloaded a version of QCTerm version 1.0 in the last several weeks and you’re running Windows 95, 98, or ME, we’ve probably screwed up your PC a little bit. Windows NT, 2000 and XP machines will be unaffected.

“The problem is not serious, only a nuisance. A user reported the problem to us this morning and it took most of the day to fully understand the nature of the problem. The problem lies with a faulty version of WinZip, Version 7.0. Somehow, the archive of to-be-zipped files created by 7.0 grabbed the entire Windows startup directory of the PC on which the archive was created. This will in turn have created a false directory of files on your PC that you will see when you go the Start/Programs function in the Start Menu.”

Version 1.01 of QCTerm is now available from aics-research.com/qcterm and is safe. If you’ve been caught by this problem, check out the Raven archives at raven.utc.edu/cgi-bin/WA.EXE?A2=ind0303C&L=hp3000-l&P=R23715 for a complete explanation and instructions on how to clean up your PC.

Some benchmarks for migrators

Relative performance numbers between systems are hard to come by so it was great to see this posting by Duane Percox: “Last year I built a Web page that contained some benchmark timings for sample COBOL code. I have updated that web page with timings from a pre-release version of KOBOL from theKompany and from the final shipping version of Fujitsu NetCOBOL for Linux. By the way, Web page is hosted on our A400 and the page is served up by QWEBS, which just happens to be written in COBOL. Here is the reference: qwebs.qss.com/cobdata.html.”

Any issues with KSAM64 files?

An interesting question and one whose answer I thought would be no. However, there is at least one little thing to be aware of. From Walter Murray, “I can tell you that I updated COBOL II to support KSAM64. If you want to read, write, or update an existing one, there should be no problem. If you’re using COBOL II to create a new indexed file and you want to force it to be KSAM64, you’ll need to use a FILE command.”

Accessing your HP 3000 from Linux

The question, “Is there a terminal emulator that runs on Linux for the HP 3000?” gets asked every other month or so and the answer is always the same: freevt3k from ftp.telamon.com. Randy Medd posted an update on the status:

“A reminder as to the status (and a little history) of freevt3k. It is not now, and never has been, a product, either of ours (formerly Telamon, now Vytek Wireless) or anyone else. freevt3k (and xhpterm) is the result of the collaboration of a loosely-knit (I’m being very generous here) group of programmers (identified in various locations in the source files), created in the distant, misty past (circa 1996.) My name is included in the list, but my contribution was minor and very much after most of the difficult work had already been done. (Meaning, please don’t ask me how the protocol or X interface works. I did most of the porting to the various Unices, but not the major design work.)

“The archive is hosted on ftp.telamon.com, but the code isn’t being actively maintained or updated. The source is located in latest.tar.Z and/or latest.tar.gz, and pre-compiled versions are available for those OS/platform combinations that we have access to here. Support is not available, except through forums such as 3000-L. (There was once a FREEVT3K-L, but I don’t know what became of it.)”

What does Filecode –491 mean?

Denis St-Amand reminded us all of a great place to find information about the meaning of various filecodes, www.robelle.com/library/smugbook/filecode.html. Stan Sieler compiled much of the information. If you want more detail, check www.allegro.com.

Whoa, Hold on There

It sometimes amazes me how many people are ready to reboot their system instead of trying to determine first if there isn’t another way. Maybe it is youth, this PC mentality. Anyway, I happened to be online when the following post came through on 3000-L: “We are currently running MPE/iX version 6.0. We will be doing an install to change the VOLUME SET to 100 percent from default 75 percent. Last time I did something like this I recall having to recreate user logons and passwords. I think I had to recreate/modify some accounts, a problem with the ‘ACCESS’ parameters. I know I had to change the access on most groups. Is all this information on my SLT tape? I used SYSGEN to create it and will be using it to boot from. Any comments or suggestions will be truly welcomed. Oh, I did have a full backup (@.@.@) using store/restore.”

I immediately responded, whoa, hold on there. If you mean changing maximum permanent space from 75 percent to 100 percent, this does not require anything other than the ALTERVOL command of VOLUTIL. It can be done online without affecting users. [Note: most people, including me, advise against raising LDEV1 to 100 percent, though actually you cannot raise it to 100 percent on LDEV 1, only close.]

Jeff Woods elaborated shortly after, “for user (i.e., non-system) volume-sets, the limits can be anything up to 100 percent. However, volumes in the system volume set (MPEXL_SYSTEM_VOLUME_SET) are limited to something less than 100 percent because some amount of transient space on the system is required. Exactly how high MPE allows depends on the size of the volume. You can use trial and error to determine how high the setting is allowed.

“But remember, MPE must be able to dynamically allocate transient space as needed. If MPE exhausts the available transient space then run-time errors will happen to applications trying to run on the system. Max-permanent at 75 percent means that 25 percent of the drive is reserved for transient space.

“Likewise, max-transient at 75 percent means that 25 percent of the drive is reserved for permanent files. But if permanent files occupy 60 percent of the drive, then transient space would be limited to the remaining 40 percent of the drive available, regardless of how much higher the max-transient limit may be. Likewise, if transient space usage is 30 percent then permanent files can only occupy the remaining 70 percent, regardless of the max-perm setting (assuming it’s not lower).”

Get used to it, there’s going to be a lot more of this in the future

As equipment gets older and older and as we neglect the maintenance habits we learned 20 years ago, we will see more postings like this: “Upon arrival this morning the console had locked up. I re-started the unit, but the SCSI drives do not seem to be powering up. The green lights flash on for a second after the power is applied, but that is it. The cooling fan does not turn either. I am able to boot, but get the following messages: LDEVS 5, 8, 4, 3, 2 are not available and FILE SYSTEM ERROR READING $STDIN (CIERR 1807)

“When I try to log on as manager.sys, I must do so HIPRI, and get the following: Couldn’t open UDC directory file, COMMAND.PUB.SYS. (CIERR 1910) If I had to guess, I would say the SCSI drives are not working. Is there a quick fix, or are all the files lost? I should add that I just inherited this system (HP937LX). It has been neglected, but running, for close to 2 years. Is it time to pull the plug?”

I am going to repeat most of the thread that followed verbatim. Tom Emerson responded, “This sounds very familiar. I’d say the power supply on the drive cabinet is either going or gone [does the fan ‘not spin’ due to being gunked up with dust and grease, or just ‘no power’?] I’m thinking that the power supply is detecting a problem and shutting down moments after powering up [hence why you see a ‘momentary flicker’].”

Tim Atwood added, “I concur. Power supply on the drive cabinet has probably gone bad. If this is an HP6000 series SCSI disc enclosure or 2 and 4 GB SCSI drives, and you are expecting to get HP hardware support on it, I would suggest you need to move very quickly. HP has already dropped support on these for many customers due to not being able to get these power supplies anymore. Even customers with long-term hardware contracts are having the HP6000 enclosures and 2/4 GB drives dropped from support March 31.

“I have also heard third-party hardware suppliers are having trouble getting these power supplies. This may or may not be true. I know the 4GB drives are getting near impossible to find. So, if it is an HP6000 series you may want to stock up on power supplies if you find them. Or take this opportunity to convert to another drive type that is supported.”

The person posted the original question came back with, “your post (Tom) gave me the courage to open the box and the design is pretty straight forward. It appears to be the power supply. As I recall now, the cooling fan that is built into the supply was making noise last week. I will shop around for a replacement. I can’t believe the amount of dust inside!”

Which prompted Denys Beauchemin to respond, “Just an FYI, the dust inside the power supply probably contributed to its early demise. It is a good idea to get a couple of cans of compressed air and clean out the fans and power supplies every once in a while. That goes for PCs, desktops, servers, your TV at home and other electronic equipment. The current is a magnet for dust bunnies and other such putrid creatures.”

Tom Emerson then reminisced, “years ago at the first shop I worked we had a Series III and a Series 48. Roughly every 3 or 6 months an HP technician would stop by our office to perform a ‘PM,’ Preventative Maintenance. Amazingly, we had very few hardware related problems with those old beasts. With the advent of the Micro 3000 and the 900 series, it seems the PC mentality of ‘maintenance by replacement’ had taken over and I was somewhat surprised one day to realize, we don’t have a tech coming out to do PM’s anymore. Sure enough, in the same span of time at my last job as I was at my first, we’ve had a few hardware related failures, including a choked-up power supply fan on a disk cabinet (which is why I recognized the symptoms listed).”

Finally, Wayne Boyer of Cal-Logic had this to say, “fixing these power supplies should run around $75 to $100. If anyone has any bad ones in need of repair, I’d be glad to furnish a firm quote. Any modular power supply like these is relatively easy to service. As to reports of a shortage of these power supplies, a quick head count here just yielded over 30 complete 6000 Series cabinets. I honestly never understand these reports of common and fairly recent (less than 10 years old) equipment being in short supply. The usual situation is that we have a glut of it and so do other equipment dealers. It is good advice to stock up on spares for older equipment though. Just because it’s available somewhere and not too expensive doesn’t mean that you can afford to be down while fussing around with getting a spare shipped in.

“The compressed air cans work but to really do a good job on blowing out computer equipment, you need to use an air compressor and strip the covers off of the equipment. We run our air compressor at 100 PSI. Note that you want to do the blasting outside! Otherwise you will get the dust all over where ever you are working. This is especially important with printers as you get paper dust, excess toner, etc. building up inside the equipment. I try and give our in-office equipment a blow out once a year or so. Good to do that whenever a system is powered down for some other reason.”

Bob J. of Ideal Computer Services added in support of Wayne, “The truth sucks. There are a couple of support companies that don’t stock spare parts. The convenient excuse when a part is needed is to claim that ‘parts are tough to get.’ Next they start looking for a source for that part. One of my former employers always pulled that crap. Unfortunately, quality companies get grouped with the bad apples. I always suggest system managers ask to visit the vendor’s local parts warehouse. The parts in their warehouse should resemble the units on support. No reason to assume the OEM has the most complete local stock either. Remember that snow job suggesting that 9x7 parts will become scarce and expensive? Different motive, but still nonsense.”

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