Hidden Value details commands and procedures in MPE (and some in Vesoft’s MPEX) that can improve your productivity with HP 3000 systems. Get a free NewsWire HP 3000 Always Online cap — submit your MPE tip directly to us here at the NewsWire. Send your tips to editor@3000newswire.com, or fax them to 512.331.3807.

How do I get the MAC address of the LAN cards on the HP 3000? I’ve got two LAN cards, one set up for the DTC (working fine) and the other for our NT network.

Glenn Mitchell replies:
LINKCONTROL @;STATUS=ALL

How do I find information on tobyte.bpbin.sys?

Christian Lheureux, Noam Koren and Andreas Schmidt reply:
From the man pages in the Posix shell on any MPE/iX system. Simply issue the commands
:run sh.hpbin.sys
(or SH if HPPXUDC is catalogued)
$ man tobyte

The documentation for both the calendar and the almanac intrinsic say the year part is the year of the century. Is this really correct or is it actually the year since 1900?

HP’s Walter Murray replies:
This is a known error in the MPE/iX Intrinsics Reference Manual. I reported it a while back in SR 4701320473 (CR JAGab46220). Bits 0:7 in the CALENDAR date format are actually the year since 1900, not the year of the century.

Where does SORT/XL create its scratch files? I have a vague memory that it defaults to the system volume set but can’t seem to discover where I got that idea. Maybe SORT/XL just creates the scratch files on the volume set where the program that calls it is running?

Stan Sieler replies:
The scratch files are opened with the formal names “HPSORTS1” and “HPSORTS2”. Thus, if you do not have file equates for them, they go into the volume set that your logon group is HOMEVSed to.

As I look at various client sites using sysinfo, in the IOMAP section I frequently see something like:

Device Adapter (10/4/24.6) —- PSEUDO
Device (10/4/24.6) —- [ldev 1 HPC2490WD DISK]
***DEV CONFIG MISMATCH, ACTUAL ID = ST15150W***

I have always assumed that these devices were configured with an ID that works okay but just doesn’t match the actual ID that could have been used. Since the devices are working, I’m not going to go in and change them, but I just found it curious. What gives?

Fred Metcalf replies:
If you locate the disk IDs in IODFAULT.PUB.SYS, you will see that all Wide/Differential SCSI disks have the same parameters. As long as you do not configure a Single-Ended SCSI disk where you really have a Wide/Differential SCSI disk, or vice versa, you can use any ID in that group. What you see is normally a result of the product number being used as the device ID (where valid in IODFAULT), while the device really has a different internal ID. Note: Watch out for an ambiguous disk ID such as ST34573WC, since it can be a Single-Ended or Wide/Differential device!

I find it confusing sometimes when looking at job output to keep track in my mind if I’m in a “COMMANDS IGNORED UNTIL MATCHING ENDIF” section or a “RESUME” section. Is there already a command in MPE that screens out the irrelevant output from a STDLIST?

HP’s Jeff Vance replies:
You can set the HPMSGFENCE variable to 8 or 16 to filter out these messages. Try HELP HPMSGFENCE.
[Editor’s note: HELP HPMSGFENCE for the enhanced functionality and SETVAR HPMSGFENCE,8 etc. only works on MPE/iX 6.0 and later. Neither HELP nor the function itself works on MPE/iX 5.5 PP7 systems.]

Is there an alternative to the program TAPETEST which used to be in the TELESUP account? It wrote a test pattern to a tape, then read it back, and reported any errors.

Stan Sieler replies:
www.allegro.com/softwa re/#TAPECHK

Is it necessary to reset all UDCs prior to issuing a PURGEACCT or PURGEUSER command, or does MPE do this for you?

Gilles Schipper, Lee Gunter and Andreas Schmidt all reply that it does.
Connie Samuel notes that if the UDC file referred to in SHOWCATALOG is in the account/group you are trying to purge, then the purge does not complete.

Gilles Schipper elaborated on the effect to command.pub.sys:
MPE does take care of it by de-activating the appropriate entry within it that corresponds to the PURGEACCT or PURGEUSER command. However, remnants of the entry do remain in the file.

There is a UDCUTIL program from the Boeing TECHXL account that can be used to clean up these remnants. One must be careful when using UDCUTIL. I would recommend creating a copy of COMMAND.PUB.SYS before executing the lnk command.

For example, if you had simply stored then purged all files from an account and so preserved its account structure, its UDC structure would also be preserved such that restoration of any still-active UDC files would render them active automatically.

If the appropriate command in UDCUTIL were to be invoked, the UDC structure of empty accounts would be removed from COMMAND.PUB.SYS. This may not be what you want to happen.

We have several archived tapes (6250bpi) from a 7980XC that I’d like to copy over to DDS media. Are there any third-party utilities available (freeware preferred), to go directly from mag tape to a cartridge, without having to restore to a local group, and store back out to the DDS drive?

Gary Biggs replies:
See TAPECOPY on the CSY Jazz Web server. 

I’m reading lines from a file into a variable FIELD then testing this field for various strings thus:

IF POS(“test val”,”!FIELD”)>0
{do code for value found}
ENDIF

The problem is that the lines being read can contain single quotes and double quotes. If I use double quotes in the POS function and the FIELD has double quotes also, I get an error. Is there some function that will change all occurrences of “ in FIELD to a single quote?

Jeff Vance replies:
Assuming FIELD is a CI variable then changing your construct to IF POS(“test value”,FIELD) > 0 should work. If you still want to replace every occurance of double quote with single quote, then setvar field repl(field,’”’,”’”) will do the trick.

I need to send out an IBM-formated tape, EBCDICOUT. The record length as specified by the recipient is 361 bytes. They report receiving 362 bytes, and cannot handle it. How can I prevent this, or what workaround can I suggest to the recipient?

Stan Sieler replies:
As far as I know, physical tape records will always be an even number of bytes on an HP 3000.

Workaround #1: Write one record per tape block, of 362 bytes, and tell them to only read 361 bytes. Their hardware/software should do the same thing ours does (if you read less data than is written): quietly ignore it.

Workaround #2: Write multiple records per tape block (an even number of records). For example, if you write two records of 361 bytes per tape block, that’s 722 bytes, which is an even number of bytes. They should be able to easily handle deblocking the data. You can probably create this trivially with: FILE TAPE;REC=-361,2,f,ASCII.

Erik Vistica adds:
Try checking the HP ESC for document MVAN000201 re: IBM Labeled Tapes Q/A. It has examples of using Stan’s workarounds with FCOPY.

Is there a way to create an SLT that includes the directory?

Giles Shipper replies:
I use the following job stream (contents of referenced file within follows):

!JOB SLTJ,MGR.SYSTEM;HIPRI
!COMMENT **********************************************
!COMMENT * ** SLTJ **
!COMMENT *
!COMMENT * THIS JOB CREATES AN SLT TAPE.
!COMMENT **********************************************
!TELLOP **SLTJ** CREATE SLT TAPE
!FILE SYSGTAPE;DEV=TAPE
!SYSGEN
TAPE STORE=”!SLTFSET.DATA.SYSTEM”
EXIT
!TELLOP **SLTJ** JOB COMPLETED SUCCESSFULLY
!EOJ

:print sltfset.data
command.pub.sys;show;directory

If you use volume sets, you should append to the sltfset record:

;onvs=MPEXL_SYSTEM_VOLUME_SET,PROD,TEST,.....

If I successfully performed: LOG logid,STOP and then shutdown and started the machine for something like replacing a tape drive — should it not be sufficient to perform LOG logid,RESTART?

Lee Gunter replies:
You don’t state the reason for your question, but if you’ve experienced a problem with a :LOG logid, RESTART, chances are you’ve run into what I believe is a bug which still exists in the user logging subsystem. This occurs when a :LOG ,STOP record is the last record written to the currently open log file and it fills the file at the same time. The next restart of this process fails because there’s no room in the current log file in which to write the restart record.

A better procedure is to include a :CHANGELOG logid (assuming you’re logging to disk and you’ve set up your logging process to CHANGE files) before you stop the logging process. This assures that your current logfile will have capacity for the restart and subsequent transaction records. We’ve set up our LOG command scripts to automate this for each logging process.

I’m trying to do a 5.0 to 5.5 MPE/iX upgrade from DDS tape and am having a serious problem. The machine hangs on the reboot from primary boot path after applying the 5.5 SLT. The last console messages were:

Leaving MOUNT_ALL_VOLUMES
The current boot command has been saved in the system master
NMLG FILE NUMBER 710 ON
10:55/23/DCC STARTUP - ENTER - version C0200007 compiled 960426
[NETWORK_DUMP_SURR] version A.00.55.001

This is where it hangs. I tried this process three separate times using two different factory SLTs and it hung at the same spot each time. What should I do to fix this?

Scott Swartzell replies:
This will sound strange, but look in your NMMGR profiles.

file nmmgrcmd=$stdinx
nmmgr
NMMGR>openconf nmconfig.pub.sys
NMMGR>pathconf dts
NMMGR>summaryconf dts profile
NMMGR>exit

This will list all of the profiles. If you have a TIO device with a record width > 510 bytes, the system will hang at DCC startup.