RAC Consulting Sponsor Message

Hidden Value details commands and procedures in MPE that can improve your productivity with HP 3000 systems. Send your tips to john@burke-consulting.com, or fax them to 512.331.3807.

Edited by John Burke

We are still on MPE/iX 5.5 PP7. Is it possible to upgrade directly to MPE/iX 7.0?

Paul Edwards replies:
No, you have to go to 6.0 or 6.5 first and then to 7.0. Why not go to 7.5 and be on the latest and last release? Note, you may have hardware issues depending on the HP 3000 you have, so be sure to check compatibility first.

Is traceroute available for MPE/iX 6.0?

Michael Berkowitz replies:
Try the patch NSTGDK3A, available from the ITRC.

Is there any way to find out if a job is scheduled BY NAME, other than brute force parsing a :showjob output? JINFO only takes job numbers. For example, if you want to see if a named job is running, you can:
:tell foo.bar Are you there?
:if cierror=1627 then
since the tell generates “J2 FOO.BAR” not interactive. Message not sent to job. (CIWARN 1627).
This little trick doesn’t work if the job exists but is scheduled, you get There was no session found that matched this ID. (CIWARN 1619)
SHOWJOB doesn’t set any return codes whether it exists or not. Is there a crafty workaround, or do you have to parse SHOWJOB output?

Gary Paveza replies:
To tell if a job is around, you could always use the JOBCNT function.

I have an older A-Class box with a flashing attention light. Does this mean anything or should I ignore it, like the oil light in my truck?

Gary Jackson replies:
Do a control B, read the ERror log (at least one record) and then the flashing light should disappear. At the prompt, type in CO to go back to console mode.

David Sims adds:
In our case it was the GSP (General Service Processor) telling us that there is an error code that had not been accessed in its store. This happens when you power down the e3000 and then power it back up again. Most of the time it is a message that can be ignored, but you will need to access the GSP’s error messages and display it to stop the annoying flashing.

How much will a DLT8000 store before it requires a second tape? Is it data dependent?

Denys Beauchemin and Jeff Woods reply:
The DLT8000 drive, using a DLT IV tape will record 40GB of native (uncompressed) data. With hardware compression enabled, this number can grow up to 80GB. Depending on the data, this number can be high or low but is a fairly reliable measure for data consisting of databases and programs and such stuff. If the data is mainly incompressible stuff, you will get less than 80GB, but not less than 40GB.

How do I know what my MPE/iX version is? When I log on I get the message: HP3000 Release: C.60.00 User Version: C.60.02

Gilles Schipper replies:
You are running MPE/iX 6.0 with PowerPatch 2.
John Burke adds:
Examine HPSWINFO.PUB.SYS to determine what patches beyond PowerPatch 2 have been applied.

Is there a version of zip/unzip that runs on MPE/iX?

Jens von Bulow replies:
Yes. You can download it from the Neil Harvey & Associates Web site at ftp.nha.co.za/hp3000/zip/

How do I configure Windows XP to use clear text passwords so I can communicate with Samba/iX?

Lane Rollins replies:
On Windows 2000 and Windows XP, Control Panel -> Administrative Tools -> Local Security Policy
In the tree inside the MMC snap in go to Security Settings -> Local Policies -> Security Options
Then change the entry “Microsoft network client: Send unencrypted password to third-party SMB servers” to enabled.
Reboot the box and you’re done.

My DLT 4000 was not detected at reboot by my HP 3000 (MPE/iX 6.0). The message was “configured but not available.” What can I do?

Giles Schipper replies:
You can try DOWNing and then UPing the device as well as power-cycling it. However, you are probably going to have to reboot.
This time, make absolutely sure that the DLT4000 is completely inactive (i.e., not in the process of loading a tape). You might have a SCSI address conflict - you can verify your actual SCSI addresses with ODE MAPPER.
To get into ODE MAPPER,
- perform a boot (from primary path),
- answer Y to interact with IPL
- type ODE at isl prompt
- type RUN MAPPER at ODE prompt
[Editor’s note: The problem turned out to be a faulty SCSI selector on the DLT and the drive had to be replaced.]

What are the NMLGxxxx log files in PUB.SYS used for, and how long they should be kept? They seem to fill up and start new files almost daily and I’ve just discovered that I’ve got over 800 of them.

Donna Garverick replies:
They are the network log files, somewhat similar to the log#### files, but strictly for network logging. As far as how long to keep them, that is up to you. Right now, because networking life is good, I’m only keeping 10 days worth of network log files.

Craig Lalley adds:
Two things could be happening here.
1) You have a problem with you network and excessive errors are being logged, in which case use NMDUMP to read the log files.
2) There are just informational messages being logged and nothing to worry about.
Personally I purge them on a regular basis. You can determine if there are errors in your network by doing :LINKCONTROL @;STATUS=all. This will let you know if the last 5 or 6 logs are worth saving.

We need to hook up two DTCs again. How can we put these DTCs on their own LAN card?

Jeff Kell replies:
Assuming you have two LAN cards, leave the TCP/IP one alone. Set up the second one’s hardware address in the DTSLINK configuration of NMMGR. You will then need the MAC addresses of the DTCs. Configure these in NMMGR, host-based DTC management, including their MAC addresses, and configure the card(s)/port(s) as necessary. Validate and exit (you probably don’t want to add all this online).

For cabling, you didn’t say which DTCs. If they are DTC48s setup by default, they will want to use the BNC connectors unless you open them up and change jumpers to use the AUI connector instead. Enable SQE heartbeat on the transceivers (DTCs and 3000). If using BNC, be sure you have both ends terminated, and use “T” connectors throughout (no shortcuts BNC-to-BNC). Reboot and pray.

Lars Appel adds:
Before resorting to reboot, I would probably first try if the dynamic configuration changes inside NMMGR (when it prompts whether to activate the changes now or later) or with DTCCNTRL FUNC=DYNCONF work as needed. If that does not help, you could still use DTCCNTRL FUNC=SHUTDOWN and FUNC=RESTART to selectively restart the DTS subsystem instead of the whole machine. Can save quite some time.

We have been concerned here with increasing backup times recently and have been having a closer look at our backup procedures as a result. We’re using plain old STORE to two DDS-3 drives (one internal, the other external). One of the things we’ve been looking at is interleaving (using the :INTERLEAVE option of STORE). Is it worthwhile to do it?

John Burke replies:
It is mostly safe, but I’ve never been a big fan of ;INTERLEAVE for the reasons you mention. DDS is just too flaky for my tastes.
I had five DDS drives and actually configured five separate, non-intersecting, backups to run simultaneously. This way, if a problem occurred on any one drive, it did not affect the other four tapes and I was at least partially backed up. The length of the full “backup” was the length of the longest individual store. With occasional adjustment, the backup could be kept reasonably in balance.

The downside to this approach is that you must be in control of system account management and (possibly) make manual adjustments to your backup scheme every time there are account/group additions. On some systems this could be daunting. However, I used the above scheme for over five years and never lost a file.