Click here for Robelle Sponsor Message

3000 NewsWire net.digest

TurboSTORE/iX, MPE/iX 5.5 and July 15: Closure

I first reported on this situation in the February issue and followed up on it in April. Since the magic date of July 15 has come and gone, it is now time to close out this thread with the truth. Unfortunately, the truth is something that corporate HP seems to have trouble with these days. And because of that, it caused many HP 3000 System Managers unnecessary worry and work. Sir Walter Scott's line immediately comes to mind: "Oh, what a tangled web we weave, when first we practice to deceive!"

Here is what happened. When HP sent out MPE/iX 5.5 it inadvertently included a full, NON-TIMEOUT version of TurboSTORE/iX 7x24 True-Online Backup (TS7x24, product number B5152AA). If you are running a nonpatched version of 5.5, you can check this out for yourself by looking at the STORE banner.

For a long time, HP was silent about the matter. Then it started putting out the story that a demo version of TS7x24 was shipped with 5.5 and would expire July 15, 1997 -- you were welcome to try out the functionality until then. Of course, as many people, including me, pointed out, the product itself did not display anything to indicate it was a demo version, like the TurboSTORE demo that came with 5.0.

A patch appeared, MPEJXD4, which, if applied (and I think it was included in any STORE related patches after some date), enforced this arbitrary July 15 date. So, if you applied this patch and were a legitimate licensee of TSII (B5151AA) or TS7x24, you would not have the functionality of TSII or TS7x24 after July 15 -- unless you updated your copy with a correct SUBSYS tape. Then came the confusing documentation about the Express 2 release. And the inevitable delay in the MR of Express 2, which meant it was June before customers could get the needed tapes.

It was not revealed that if you had not applied MPEJXD4 and were a legitimate user of TSII or TS7x24, you did not need to rush to apply Express 2 or perform the alternate procedure to update TurboSTORE. July 15 would come and go without causing a blip on your operations. However, if you had applied MPEJXD4, then you had better get your Express 2 tapes and update TurboSTORE, or else you would have a disaster.

HP decided it could not trust its customers with the truth and to act ethically on that knowledge. So it created a situation with patch MPEJXD4 (the tangled Web) and deliberately confusing documentation and articles (the deception) that caused panic and could have -- though I do not have any evidence that this occurred -- caused backup failures and down time at some of its largest customers. At the very least, as evidenced by postings to the Internet, it caused considerable, unnecessary concern among many HP 3000 System Managers.

Here is what should have been done:

• Admit the goof at the earliest possible time;

• Propose that non-licensed users of TS7x24 consider it a demo with a demo period expiring, say, July 15, or 90 days after installing MPE/iX 5.5;

• Publish the procedure to disable TS7x24 in favor of FOS STORE for those customers without a license to TSII or TS7x24 and who do not want to use the demo version of TS7x24;

• Create a plan, and publish it well in advance, to resync the products with customer license beginning with the first Express 5.5 update.

Host-based inbound Telnet:
a reality, but how well does it work?

Host-based inbound Telnet is still relatively new: MPE/iX 5.5 PP1 plus patches. This thread was started by a series of questions: "Is anyone utilizing inbound telnet on MPE/iX 5.5? If so, how many users do you have coming in and what system are you running? Have you run into any limitations?"

Few people have much experience and there were several responses such as: "It works for me. Of course, I'm the only one using it, so performance is great. Just make sure you have all current Telnet patches installed or you could be asking for crashes." Or, "I've had inbound telnet up and running for two or three weeks now without any problems. There are only a few users at present, and the heaviest load was five concurrent inbound telnet sessions. I haven't heard of any problems and haven't had any aborts or system failures. The environment? MPE/iX 5.5 with Power Patch 1 and a few patches on top of that. Configuration? Mike Tilford of HP posted a quick guide to installing inbound telnet a while back (see below) and I used it to set up inbound telnet on the system here."

Only one person reported any significant use: "During our last major fee payment period, our DTC TAC became swamped at the limit of 40 connections. As a bold quick fix, I reset the DNS name of the TAC to point to the 3000's IP (after having only casually tested inbound telnet for myself). The remainder of the day was fine. The next morning was more interesting, as all telnet was now coming in host-based. We had approximately 48 telnet sessions at one point on a 960. Response was a bit sluggish, but it always is during these peak demand times, and I can't point a finger at telnet as the culprit. Measuring telnet overhead is a mystery though. inetd spawns a JSMAIN process directly so there is no equivalent of a VTSERVER process to account for the CPU time used processing the transport layer. Is the telnet overhead charged to the JSMAIN, the user process or system overhead?" [Editor's note: I guess only the Shadow knows.]

Mike Tilford of HP posted, "I had success configuring the in-bound Telnet capability of MPE/iX version 5.5 using Powerpatch 1 so I thought that I would share what I did to accomplish the task. Unfortunately, the customer did not have a printer connected to the 3000 yet, so my 'sharing' is only as good as my notes and memory. The Response Center can also help with this, if you run into difficulty.

"After Powerpatch 1 is installed and the system is again operational, a few final steps need to be performed for in-bound telnet to function properly:

€ Logon to MANAGER.SYS,NET

€ Do a :LISTF SERVICES,2