Q: What's wrong here?

LISTF @QPLIMIT@.@,2
^
File name is more than eight characters long. (CIERR 532)

Would this not be considered an implementation flaw? After all, does not @ match nothing as well?

A: Jon Diercks replied:

My impression is that :LISTF is "mature" and not likely to be modified/enhanced much. LISTFILE works as you would expect:

:listfile @qplimit@.@

Non-existent MPE named file "@QPLIMIT@.@.SYS". (CIWARN 920)

Q: I'm having problems with a job I'm testing. It appears the job breaks out of the "while" loop and aborts. Why would a job treat the 'while' statement any different?

The command works well with the first time thru the loop, but then exits the loop and aborts. Can anyone see what I'm doing wrong here? This is the "while" section of the job:

: while setvar(rec_cntr,rec_cntr+1) <= !jabrt
: print jaborts;start=!rec_cntr;end=!rec_cntr > recout
: input recin < recout
: setvar msg word('!recin',"#",2)
: hpemail
JOB ABORTED
!msg
602nnnnnnn@mobile.att.net
: pause 5
: setvar page_alert true
: endwhile

A: Jeff Vance replied:

You cannot provide "inline" data inside a CI while loop. Try creating a temp input file for the HPEMAIL script, e.g.:

...
echo JOB ABORTED >mailin
echo !msg >>mailin
echo 602nnnn... >>mailin
hpemail <mailin
...

Not a question, but a tip from Ken Sletten:

In preparing for an update from MPE 5.5 to 6.5 EXP2, I noticed that VSTORE "FILES VERIFIED" versus STORE "FILES STORED" were NOT the same: STORE says it wrote one more file to tape than VSTORE says it verified.

Turns out it's a known problem: SR 5003337162/JAGab20221. STORE creates the file "/_HFSACCT/_HFSGRP/HFSMAP" on the STORE tape. It is a directory file for all HFS files on the tape. When STORE statistics are compiled, STORE includes this file in the "files stored" total. But VSTORE does NOT consider /_HFSACCT/_HFSGRP/HFSMAP to be a "real" file and does

not include it in the "files verified" count.

Q: I have received documentation for MPE/iX 6.5 PowerPatch 2. It mentions that the last date to request this is 2/28/2001. Is the tape I receive now the same as one I order four weeks from now? Or are additional patches bundled in the PowerPatch as time goes on?

A: Jon Cohen replied:

PowerPatch 2 is a collection of Generally Released patches that are integrated & certified together. If we were to add more patches to the mix, we would call the result PowerPatch 3. In general, the contents of a PowerPatch do not change after its original shipment date.

Q: Does anybody know how to run the Software Inventory Report, psswinvp.pred.sys, under 6.5? Is there a replacement? I have always executed it as follows:

:run psswinvp.pred.sys;xl='xl.pred.sys,xl.diag.sys'

It was my trusted friend on MPE/iX 5.5, but since upgrading two of our three boxes from to MPE/iX 6.5, all I get is the program banner.

A: Steve Pittenger replied:

There seems to have been a change in MPE/iX 6.5. I had the same issue when I upgraded to MPE/iX 6.5. Here is the command that my ASE gave me:

RUN PSSWINVP.PRED.SYS;XL='XL.PRED.SYS,DIAGXL.PRED.SYS'

[Editor's note: this command is contained in the command file psirpt.pred.sys.]

Q: Are add-on NIC cards for 9x7's and 9x8's full-height like a MIO card or half-height like a single-ended SCSI card?

A: Doug Werth replied:

The add-on LANIC is a half-height card.

Q: Correct me if I'm wrong, but if my memory serves me right all one has to do to change the IP address on a Native Mode spooled printer is:

STOPSPOOL ldev

Edit NPCONFIG.PUB.SYS

STARTSPOOL ldev

Or is there something more complex than that?

A: From several people:

No, it is that simple in general.

If you use DNS and name your printers, then, as James Alexander pointed out, you need to clear the cache entry with nettool.net.sys. [Editor's note: if the printer is changing subnets, then you may need to change the gateway on the printer's JetDirect interface.]

Q: Is there a CI command that is the inverse of

SHOWDEV CLASS

In other words is there a simple command available at the CI that, when given an ldev #, will return all of the logical device classes that it belongs to?

A: Larry Simonsen replied:

Not exactly a command, but here is a script using sysinfo.prvxl.telesup

parm device
echo io !device>sprompt
echo exit>>sprompt
sysinfo.prvxl.telesup <sprompt

Lars Appel suggested using the implied "info" string to make it even simpler:

:sysinfo.prvxl.telesup "io !device"

[Editor's note: There was a period in the MPE/iX 5.5 time frame when sysinfo lost the ability to display classes. I recall complaining about it, but since sysinfo is officially "unsupported", I did not get very far. It appears to have been fixed with version A.04.00 which was delivered as part of MPE/iX 6.0 PowerPatch 1.]

Q: I'm trying to set up a udc that either calls a command file or calls itself. Example,

spoolf dfid
tellop user !hpuser.!hpaccount is modifying !dfid
spoolf !dfid;pri=1
*****

If I use 'spoolf' as the name of the udc command then the udc begins an infinite loop. What's the work around, if any?

A: Doug Werth replied:

Add OPTION NORECURSION to your UDC to prevent it from executing itself.

Jeff Vance added:

A UDC can easily call a command file, either my using the file's fully qualified name, or letting the name be qualified via the HPPATH variable. If the command file has the same name as a UDC command or a built-in CI command (consider changing the filename)then use the XEQ CI command.

A UDC can call itself by using OPTION RECURSION in the UDC header or as a command in the UDC body.

Q: We are trying to setup a command file that will read a record into a system variable and then echo it into a file using CIOR. The problem is that the records could (and do) contain the characters > and <. The command interpreter is evaluating these as part of command IO redirection and failing on "file not found" or "illegal filename" errors. How can we override this?

A: Keven Miller and Chris Goodey replied:

If you put an "!" in front of the ">" and ">" the characters will pass through, as in this example:

SETVAR LINE REPL ( LINE, ">", "!>" )
SETVAR LINE REPL ( LINE, "<", "!<" )
echo !LINE

Q: We had the 'pleasure' (not) of having an AUI transceiver fall off the back of one of our 3000s the other day. In this case, the transceiver doesn't have the little "locking bracket thingy" and so it's simply hanging on the AUI port. Our tech tried to tell me that he can't buy a transceiver that fits/screws on to a MFIO card. I can't quite believe this is so. Anyone want to disagree? Anyone want to give me a part number?

A: Rick Gilligan replied:

I highly recommend the HP 28685B EtherTwist Transceiver. It not only comes with the "locking bracket thingy", but also is held tightly in place by it. We had tried a couple of other brands of transceiver with the little posts, but none were held in place tightly, except for the 28685B.

Denis St-Amand added:

The part you need (for the "locking bracket thingy") is 5062-3351, an "AUI retainer".

Bob J. and Stan Sieler added:

Most transceivers block the retaining clip so you can't lock the clip. To solve this get transceivers with pigtail cables or get a short AUI extension cable.

Stan further added (crediting Steve Cooper) that if you need a solution RIGHT NOW, slightly crimp the "shell" of the AUI connector so the connector will hold on to the card better.

TIP: (This came as the result of a fairly long thread that dealt with a network printing problem.) NPCONFIG.PUB.SYS must be an unnumbered file.

Q: How can I determine how much memory is on each of my systems?

A: Larry Simonsen replied:

The (free) ramusage utility at www.allegro.com will tell you memory size.

Barry Lake added:

If you have Glance/iX, then the "Memory Detail" page will give you total memory. Also, psconfig.pred.sys will show memory detail.

Wirt Atmar noted that the initial banner for sysinfo.prvxl.telesup display total memory information.

And, finally, Andreas Schmidt noted that on 6.0 and earlier systems, the following works:

SYSDIAG
SYSMAP
MEMMAP

Q: I just updated our development system from 5.5 to 6.0 pp2. Doing a SHOWME yields the following line:

RELEASE: C.60.00 MPE/iX HP31900 C.16.01 USER VERSION: C.60.00

Shouldn't the "User Version" be C.60.02?

A: Gilles Schipper replied:

I believe the USER VERSION: C.60.nn will vary according to which tool you used to apply the powerpatch.

If you use autoinst, nn will equal 00.

If you use patchix, it will show 02.

In both cases, HPSWINFO.PUB.SYS should accurately reflect the patches applied and powerpath level.

Several people noted that you can alter the value of "User Version" in SYSGEN, but you must create an SLT and update from tape for the new value to take effect.

Q: It's been a long time since I monitored console messages but I know that DTC downloads used to be logged to the console. Yesterday I moved the console to my terminal (Reflection VT session) and rebooted a DTC but there was no message indicating that a download took place.

Is it possible that console logging of DTC downloads has been turned off on my machine? I'm running 6.0 PP1.

A: Kevin Miller and Jim Phillips both replied that the DTC download messages go to the physical console (ldev 20).

Q: DBGET mode 8 will return the primary entry for a given argument. Is there a way to return all the synonym/secondary entries for previously returned primary? Also, is the hash algorithm that IMAGE uses for masters published somewhere?

A: Denys Beauchemin replied:

Yes, simply follow the chain using DBGET mode 5 on the master.

The algorithm is published in the Image Handbook.

Q: I can't find an HP variable which holds the name of the jobq that a job is in. I would expect a name like HPJOBQ to be appropriate. Also, I want to set the JOBPRI for specific queues but the JOBPRI command does not appear to support the ;JOBQ= syntax of other job commands like SHOWJOB and ALTJOB.

A: Larry Simonsen, Michael Berkowitz and Jeff Vance all replied to the first question:

There currently is no HP variable, but you can easily create one with the "jinfo" function, as in

setvar myjobQ JINFO('0','jobq') # zero means "me"

Jeff then responded to the second question:

By design, job queues only support two properties: name, limit. This may be extended over time, but that's all for now. A reason for this was to not compete with existing 3rd party solutions in this area.

Q: I'm running into a problem with Patch/iX. I get to the 'Create the Tape' step, and the process seems to hang at the point of 'Modifying the Staged SL in SL.INSTALL.SYS'.

I'm on 5.5, with Patch/ix version B.00.17. Any ideas?

A: Dennis Heidner replied:

Patch/iX will act really strange if the LP isn't spooled (STARTSPOOL) and batch stream device (STREAM 10) hasn't been issued. I believe later versions (MPE/iX 6+) check for these conditions, but the 5.5 version didn't.

Q: I am on MPE/iX 6.0, trying to add a new network printer. I was under the impression that you could have more than one printer with the same name, e.g. 'LP'. However, when I try to add it in IOCONFIG, I get an error. Any ideas?

A: John Burke replied:

If the classname already exists, then you do not want to add it in IOCONFIG, but rather modify it by "adding" an ldev: mc foo aldev=n

You should also stop the spooler before "adding" the ldev and then restart it upon exiting ioconfig (it will probably start automatically).

Q: We have a HP3000 918LX that is the last remaining Bisync holdout. I would like to put an RJ45 transceiver on the back of the console card and do away with the Bisync connection. I have checked and there definitely is an AUI port on the console card. How do I activate it?

A: Larry Simonsen, Gary Jackson and Bob J. all replied:

You can use the AUI lan port on the MFIO card with a transceiver to convert to a RJ45/10baseT connection. Just connecting the transceiver is not sufficient since the card has a set of jumpers that determine which lan port to use. You have to shut down the system, power it off, remove the MFIO card and re-position the jumper.

Tom Emerson added a note of caution:

Don’t forget that when you do this, be sure to turn ON the "sqe heartbeat" on the transceiver.

Q: Is it possible to alter the samba configuration file and reload samba without resorting to aborting and re-streaming the two background jobs. (I run them separately from the JINET job).

A: John Burke and Lars Appel replied:

It probably "depends". I know you can add or modify shares without stopping and restarting jsmb. At least this is one instance where you can try making the change to see if it takes without worrying much that it will harm anything.

Lars added:

My personal, and in no way authoritative, "rule of thumb" is "changes to share definitions just need a reconnect, changes to global-style parameters better have a job restart".

However, as Windows and SMBD tend to do some caching on some of their data structures, it sometimes helps to use smbstatus to look up a client's associated SMBD process and send it a kill signal in the shell, to force the next client access re-launch a fresh child.