net.digest June 2002

June 2002

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

The merger is now a done deal and my prediction stands: “Regardless of the outcome, everyone involved will come out of this diminished.” This will doubtless be the most watched post-merger integration in history, with both pro- and anti-merger groups anxiously waiting to say, “I told you so.”

One of the first casualties of the merger was CSY itself. What is left of CSY now reports to HP’s server marketing group. I’m reminded of all the CSY staffers saying there is no need to rush into decisions; after all, “it is business as usual until November 1, 2003.” I was worried CSY would implode, but I never figured it would happen so soon. One has to wonder how the remnants of CSY will fare in a marketing group devoted to HP-UX.

Will solutions other than migrating to HP-UX get a fair shake? Will HP-UX marketing be thoroughly distracted? It might be circling the wagons after the new president of the “New HP” said on his first day on the job, “You are going to see Windows and Linux absolutely eviscerate the midrange proprietary Unix.” HP’s head of marketing for HP-UX, Mark Hudson, was then prompted to blanket the Internet with a “HP-UX is strategically important to HP” missive. Does that sound familiar, MPE users?

With the merger settled, 3000-L got back to the usual array of off-topic and wildly off-topic threads. One thing you can count on is that any comment dealing with religion or politics will generate a seemingly endless thread where people expound their positions over and over and no minds are changed. This month, we actually had a thread that combined both.

It started out with a comment about global warming, briefly stayed on the topic of scientific proof or disproof and then veered wildly off into discussions about politics and religion. Other notable threads discussed the 75th anniversary of Lindberg’s arrival in Paris, the fair market value for a 9x7 and various tax implications, and “How IBM got it right with the iSeries.”

As always, I would like to hear from readers of net.digest and Hidden Value. 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, or if you’ve got an idea for something you think I missed, let me know. You can reach me at john@burke-consulting.com.

Danger, danger, Will Robinson

I’ve been a big supporter of “NSCONTROL KILLSESS=#S...” for aborting VT sessions since it was first introduced. I had evidence that under certain circumstances a simple ABORTJOB could leave a hung session laying around. The JMAT entry was gone, but the vtserver process still existed and at that point you were stuck with a reboot to clear it, whereas “KILLSESS” seemed to always work. For years now, I’ve been using without problems a command file to abort all or selected users that relies on “KILLSESS” for VT sessions. In fact, I was among those who influenced Jeff Vance to include KILLSESS in his ABORTJOB command file (on Jazz).

Apparently, some users of the ABORTJOB command file on MPE/iX 6.5 and MPE/iX 7.0 are encountering hung systems or an SA614. As is often the case with things networking, James Hofmeister provided the answers:

A SA614 can be caused by “NSCONTROL KILLSESS=#S...”. This is fixed on 6.5, 7.0 and in base 7.5: SR 8606185990 — SA614 SS 143 6.5, $b_loop, sk_log_error+$210 from killsess of vtserver.

The fix is available in General Release patches: NSTGDB1 for C.65.00 and NSTGDB2 for C.70.00. These patches should be available to download from the HP-ITRC. So, my advice to you (advice I intend to follow myself) is that if you are on MPE/iX 6.5 or MPE/iX 7.0 and have not applied the appropriate patch from above, do NOT use “NSCONTROL KILLSESS=#S...”.

A new old FTP trick

With the introduction of MPE/iX 6.0, FTP was changed significantly. Among the changes, when FTP became part of inetd, was that connecting to an HP 3000 using FTP no longer launched a session. The impact of this is that if you depended upon logon UDCs for security, etc., you no longer had this capability.

A user asked 3000-L if there was a way to prevent people from FTPing into a “user.account” that did not have passwords. James Hofmeister posted an answer about a feature that has been part of FTP for some time (from before FTP was rolled into inetd), but I expect many of us missed it because it was not relevant at the time. Here is James’ response:

One way to accomplish this was an enhancement put into FTP/iX many moons ago. A customer did not use MPE passwords, but had VEsoft or application passwords and did not want FTP users being able to PUT, GET, DEL files on a ‘user.account’ with no MPE passwords. The user was able to use this enhancement to secure these non-password logons from FTP/iX.

From FTPDOC.ARPA.SYS, the ‘PASSWORD = ON’’ SETPARMS.ARPA.SYS configuration option is a security enhancement added to the FTP Server ‘FTPSRVR’ to support the ability to restrict the establishment of FTP connections to MPE USER.ACCOUNT’s where at least ‘one’ password (a ‘USER’ password or an ‘ACCOUNT’ password) must exist.”

James adds to keep in mind this feature is enabled system-wide. For more details, see FTPDOC.ARPA.SYS, which is included with the current GR FTP patches for 6.0, 6.5 and 7.0.

HP Software Media is updated

A month or two ago, I wrote about the tale of woe one user faced when his new N-class HP 3000 came with only one tape drive, a DDS-4. The problem was that all HP software was distributed on the 60-meter DDS-1 tapes, and the DDS-4 drives cannot read these tapes. He went back and forth with the Response Center and, eventually, HP actually loaned him a DDS-3 tape drive so he could install the software. He finished up his posting with a plea for HP to fix this situation.

It appears HP was listening. Many of us got a notice from HP this month: “Due to new format DDS media devices and the obsolescence of the DDS-1 devices, HP is discontinuing software distribution on DDS-1 media format, effective August 1, 2002.” New tapes will be distributed in DDS-2 format.

Cool. Jeff Woods posted on 3000-L this further clarification of DDS reading and recording issues, something that you might want to pin on your bulletin board. He began with a table (below) showing DDS media compatibility.

(Native; ignores compression)

Recording Mode

Drive DDS-1 DDS-2 DDS-3 DDS-4

DDS-0 <60m 60 m

DDS-1 <60m 60 m 90m

DDS-DC <60m 60 m 90m

DDS-2 <60m 60 m 90m 120m

DDS-3 <60m 60m 90m 120m 125m

DDS-4 120m 125m 150m

Capacity 1.3 Gb 1.3 Gb 2 Gb 4 Gb 12Gb 20Gb

The recording mode used depends on the media inserted rather than on the drive type. This allows, for example, a DDS-3 drive to write on DDS-2 media in a manner readable by DDS-2 drives. Note that the data density is a constant within each recording format. Higher recording formats achieve progressively greater density. DDS-4 had to sacrifice DDS-1 compatibility due to the disparity of recording densities and techniques.

Note also that mixing significantly different media lengths on a single drive is discouraged because the longer tapes are much thinner, which causes the tape to take on a different shape as it’s held taut over the tape head which in turn can cause the tape head to wear differently. Ideally, you should pick one tape length (really tape media thickness and pliability) and use it exclusively in that drive. Trying to use both 60m or shorter and 120m or higher tapes on a frequent basis in a single drive is a really good way to make the drive develop reliability problems.

HP Eloquence — the real deal?

At the Solutions Symposium in April, CSY, (oops, there is no entity called CSY anymore — but there was then) announced support for HP Eloquence (which runs on HP-UX, Linux and Windows) as suitable for those migrating to another platform but still needing an IMAGE look-alike. The database’s developer and head of Marxmeier Software AG Michael Marxmeier spoke at the Symposium and presented a good case for HP Eloquence. One quote stood out as relevant to TurboIMAGE users: “Anything that TurboIMAGE does, Eloquence should be able to do, or it is a bug.”

The SIG IMAGE IMAGE/SQL Advisory Committee (ISAC) has also made a qualified endorsement of HP Eloquence. The only question appears to be how high Eloquence will scale in practice and still maintain adequate performance. Marxmeier, when asked about scalability, reported that there were no theoretical limits beyond those posed by the operating system, but admitted that the largest user of the software probably had fewer than 500 concurrent users. The IMAGE lab reviewed HP Eloquence prior to CSY’s endorsement. Perhaps as a last act they could due some stress testing.

A number of tools vendors have announced some subset of their products already work or will shortly work with HP Eloquence. They include (listed in no particular order and probably incomplete) Advanced Network Systems, Minisoft, MB Foster, DISC, Speedware, Cognos, Robelle, VitalSoft and AICS.

To date, there has been little HP 3000 user experience with HP Eloquence made public. However, Duane Percox of 3000 ISV QSS did make the following comment from an application developer’s point of view: “I have met with the principals for HP Eloquence and believe they are committed to this database. I would have no problem committing to it, and in fact we are going to use it with a product we have that is Speedware-based which we are porting to HP-UX.”

It’s already beginning

A customer noted on 3000-L that HP is discontinuing sales of processor upgrades (#A3616A) for the HP 3000 997 servers effective June 1. He asked, “Will I be able to purchase these upgrades from a reseller in the future?” Wayne Boyer had what I thought was a good response, which I’m going to quote here. HP, are you listening?

“Processors like any other piece of hardware are available and will continue to be available on the used market. A processor upgrade implies a license change though. The problem will be how to properly tell the system that additional hardware is installed without using HP’s SSCONFIG software. I’ve posted a few comments before related to SSCONFIG — this is now a real life example of licensing problems that will begin to occur. Simply buying and installing additional processors will not change the firmware settings in the 997.

“A question for those who know more than me: Will the system even sense and use additional processors if the firmware settings are NOT changed? If so, then this is only a licensing issue. (HP: what is your position on this?) If the hardware will not sense and use the additional processor(s) without a firmware change, then this is a real performance/upgrade limitation. OpenMPE Folks: Any news from HP on the SSCONFIG issue yet? Interex folks: Wouldn’t this be an important ‘advocacy’ issue for Interex to take up with HP?”

Let me second Wayne’s remarks and then add my two cents. With the recent implosion of CSY, I’m becoming increasingly concerned that November 1, 2003 will come and go and that there will still be significant questions left unanswered on the table. What we need now is for someone at HP to come forward and make a statement of direction. Waiting until this year’s HP World is too late.

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