Hidden Value February 1998

Click here for TSG sponsor message

Hidden Value details commands and procedures in MPE 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 rseybold@zilker.net, or fax them to 512-657-3264.

Compiled by John Burke

It is my understanding that after bringing the system down with CTRL-A shutdown, that I am supposed to press CTRL-B and get a “CM” prompt, at which time I either enter TC or RS. My problem is that I never get the CM prompt. The console does not respond at all. To restart my 928RX, I have to power it down and back up, which I don’t like to do. Am I misunderstanding something?

Gilles Schipper and Gary Biggs reply:
There is a black switch next to the UPS port at the back of the CPU. Switch it to the “SERVICE” position to permit console access to the CTRL-B action. The “NORMAL” position will disable console access to the CMP (control and maintenance processor). In this position, HP’s equivalent of CNTL-ALT-DEL is disabled. Power cycling becomes the ‘normal’ way to reboot.

Jeff Kell adds:
Some other systems have a key switch. [Editor’s note: the 9x9s have a key switch on the front which must be turned to “SERVICE” (go figure) to get a CTRL-B prompt.]

What is the maximum number of items allowed in an IMAGE detail dataset?

Jerry Fochtman and Mike Berkowitz reply:
TurboIMAGE raised this limit to 255 ‘fields’ per dataset. However, this is further constrained by TurboIMAGE’s ‘block size’ limit of 2560 half-words, which is the maximum size an IMAGE block can be and contain all the pointers for the entry(ies) along with the block’s entry bit map. Depending upon the size of the field items, one can define < 255 fields in a set, yet not be able to build the set due to the maximum IMAGE block limit.

Vikram Kumar adds:
Users are also limited by the total number of 1023 fields in the entire database.

I know that every connection on a coax (10 base-2) network must be terminated. But can I have an unconnected “T” on a 10 base-2 network segment?

The consensus is that 99 percent of the time the network will perform fine with one or more unconnected T’s. In fact, some people even do it on purpose so they can have someplace to plug in test boxes. However, there is also strong sentiment for junking all coax in favor of untwisted pair 10-Base-T. Your mileage may vary.

When I execute an MPE IF statement, MPE generates output to $STDLIST showing the result of the IF statement, the commands being skipped, and so on. Is there a way to suppress the output from the IF statement and associated statements like ELSE and ENDIF?

Within MPE now, it is a kludge at best of nested command files redirected to $NULL. In 6.0 you will be able to set the HPMSGFENCE variable to suppress the echoing of commands in FALSE command blocks (e.g. if false then ..won’t show the skipped commands). You will also be able to set HPMSGFENCE to suppress the “*** EXPRESSION TRUE/FALSE” “*** RESUME EXECUTION OF COMMANDS” etc. messages related to IF, ELSE, ELSEIF, ENDIF, WHILE and ENDWHILE. HP says no 5.5-based patch is available for this enhancement.

Mark Bixby adds:
You could switch now to using the sh.hpbin.sys POSIX if statement, which never generates any extra output.

What are the limits for session JCWs and VARs?

HP’s Jeff Vance and Scott McClellan reply:
Here is a way to find out:
deletevar @ setvar i 0 while true do setvar ![chr((i mod 26)+ord(‘A’))]!i, !i setvar i,i+1 endwhileThe “setvar ![chr((i mod 26)+ord(‘A’))]!i, !i” line creates a variable starting with a letter from A-Z with an integer value.

On my 5.5 system I can create 10,516 simple variables. Note that this represents a best-case scenario, since I chose short variable names and integer values. The number decreases with long variable names and long (string) values. The example below creates variables with 21-25-byte long names and a 20-byte-long value:
deletevar @ setvar i 0 while true setvar ![rpt(chr((i mod 26)+ord(‘A’)),20)]!i,rpt(‘a’,20) setvar i,i+1 endwhileand is able to create 8,440 variables.

So I take my Windows 95 PC, which connects to our network fine, and plug it into a hub on a newly set up Series 917. I fire up Reflection, try to connect to the IP address, and then wait for up to five minutes to get a login prompt. Once I get connected it seems fine. What kind of things should I consider to get that login faster?

Denys Beauchemin and Melvin Rees reply:
The problem is one of name resolution. The PC is going through a series of name queries before it tries the name as an IP address. This can be incredibly slow. It could be trying HOST, PROBE, DNS, and finally the IP address. You should verify that you have the HP 3000’s name in the HOSTS file found in the directory of your TCP/IP stack on your PC. Either \windows if you use Win95 TCP/IP, or \reflection (or the appropriate one) if you use WRQ’s stack. The time-out value is 255 seconds before a different attempt is made at finding the host, hence the long delay. Once the resolution is made, everything will indeed go fine.

I’m looking for a means of displaying a warning against unauthorized system access prior to logging on. Our corporate security folks are probably going to mandate such a message be displayed (four to six 80-character lines) on all our HP 3000 systems, and they want to have it displayed prior to the system logon prompt. I know that the ‘logonprompt’ is configurable via SYSGEN, but I don’t think it’s capable of such a message (more than 255 characters). Does any other solution exist?

Mike Hornsby replies:
On our web site we have a utility called LPRESET that does what you are looking for. It allows you to display a short warning and a prompt.

Is there a free (or nearly free) way to send SMTP mail from an HP 3000 to either a Unix Box or NT SMTP gateway?

Stan Sieler replies:
Yes. Telamon has a mail program that does it free. It works well. Check: ftp://ftp.telamon.com/dist/mail.nm or ftp://ftp.telamon.com/dist/mail.nm.wrq. Note: The Telamon MAIL program must be in a group with PM capability.

What does the “:C” or “:S” mean in the VOLUME RESTRICTION column of a store listing? What manual is it found in?

Michael Hornsby and Wolfgang Kinscher reply:
S is Volume set restriction, C is Volume class restriction. The NEWCLASS command in VOLUTIL lets you build subgroups of disks in a volume set; for example, NEWCLASS BGD:VC1 MEMBER1,MEMBER2.

This can be useful if you want to force specific files to specific members within a volume set. This is then specified with the DEV= parameter of the BUILD command. When you BUILD xyzzy;dev=VC1, the file is then forced to MEMBER1 or MEMBER2 of the BGD volume set. This is all documented in the Volume Management Reference Manual.

I have a problem with an account while running in the shell. When I do ‘sh.hpbin.sys -L’ I get:
tput: Unknown terminal “hp2392a”. tabs: Unknown terminal “hp2392a”.After this, from the shell/ix prompt, I can’t use more, or vi, among others. I get the same kind of error, about unknown terminal. This all works fine from several other accounts, but not from this one. What do I do to make it work?

Richard Gambrell replies:
Sounds like a permissions problem. As root (oops, I means manager.sys), go into the shell and do:
chmod -R a+rx /usr

This allows all to read and execute everything under /usr, which includes the terminfo files that are needed by vi to control the terminal.

Am I using LISTF wrong? When I issue the LISTF,2 command, in the middle of the resulting report I get a line that says “Cannot access file “BAT4IN” due to lock contention. (CIWARN 9169).”

HP’s Jeff Vance replies:
You must be on Express 3 or 4. LISTF “access” functionality was introduced in Express 3. The short answer is this is a known problem with Express 3. It has been patched (MPEKX18) and is also fixed on 6.0. If you re-issue the LISTF it will likely work, although I know this doesn’t help a bit for LISTFs in jobs. The KX18 patch removes the lock contention problem, fixes a bug in LISTF error handling, and reports accurate EOF for LISTF of files being accessed.

How do I point network printer configurations to specific ports on (external) multi-port JetDirect (or equivalent) boxes?

Gilles Schipper replies:
You need to add the tcp_port_number option, in NPCONFIG, as follows:

(network_address = 128.250.232.40 tcp_port_number = 9100) # for port 1 (network_address = 128.250.232.40 tcp_port_number = 9101) # for port 2 (network_address = 128.250.232.40 tcp_port_number = 9102) # for port 3
Please note that everything on each line after and including the “#” represents a comment.

What are the negative values used in FCRANGE in the STORE command to get only database files? For example, to store only databases that start with the letters ONRL@ you would say:
:STORE ONRL@;SHOW;FCRANGE=-n/-m;DIRECTORY;.........etc.

[Ed. note: Several people posted, and many more, myself included, probably readily agreed that the answer is -401 and -400. Stan Sieler set us all straight:]

Someone just posted -400 & -401, which is the old answer – but only for some databases. Here’s the list of -4xx filecodes I have:

! # Mnemonic ! Usage First seen? !–- ––– ! –––––––––––––&# 150;– –––––- ! HP defined privileged filecodes (SS mnemonics)... -400, “iroot “, ! image root -401, “idset “, ! image dataset -402, “irdba “, ! image remote database access -403, “idump “, ! image i-file (and j-file) -404, “ilr “, ! image intrinsic level recovery -405, “iglob “, ! TurboIMAGE/XL control block file -406, “idrb0 “, ! TurboIMAGE/XL DRB’00 file -407, “igtx “, ! TurboIMAGE/XL GTX file -408, “iccf “, ! TurboIMAGE/XL Jumbo chunk control file -409, “icdf “, ! TurboIMAGE/XL Jumbo chunk data file -410, “itpir “, ! TurboIMAGE/XL TPI “root” file (was: omni1) -411, “itpid “, ! TurboIMAGE/XL TPU “data” file (was: omni2) -412, “iidx “, ! TurboIMAGE/iX ..idx file -420, “irecv “, ! DBRECOV restart file -421, “chgcf “, ! DBCHANGE control file ! SS 940426... (from Goetz Neumann, HP Germany) -430, “tc “, ! TC file (IMAGE root name + “TC”) links root -> DBE -431, “atc “, ! ATC file (links a DBE -> IMAGE roots) -450, “dausf “, ! DAUSF is the NETUsers config info (SC 940930) -451, “daalf “, ! ?? created by NETUTIL (SC 941001) -491, “dbe “, ! DBCORE database environment (“DBDB” in first 4 bytes) ! ALLBASE_DBCON_FILECODE -493, “ptpt “, ! DBCORE Chunkset (“PTPT” in first 4 bytes) -497, “lflf “, ! ALLBASE (DBCORE) Logfile (“LFLF” in first 4 bytes)Is there a manufacturer of a JetDirect-compatible print server (i.e., one that will work with MPE/iX 5.5’s network printing) that has one or more serial ports? As far as I know, HP does not make a JetDirect Ex box with a serial port. I would like to eliminate a number of remote DTCs (and the bridging they require) whose only function in life these days is driving serial printers.

Life’s tough. Either use a JetDirect Ex with parallel-to-serial converters, or use a third-party print server with serial ports and third-party software to drive it.




		Click here for TSG sponsor message

		Hidden Value details commands and procedures in MPE 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 rseybold@zilker.net, or fax them to 512-657-3264. Compiled by John Burke It is my understanding that after bringing the system down with CTRL-A shutdown, that I am supposed to press CTRL-B and get a “CM” prompt, at which time I either enter TC or RS. My problem is that I never get the CM prompt. The console does not respond at all. To restart my 928RX, I have to power it down and back up, which I don’t like to do. Am I misunderstanding something? Gilles Schipper and Gary Biggs reply: There is a black switch next to the UPS port at the back of the CPU. Switch it to the “SERVICE” position to permit console access to the CTRL-B action. The “NORMAL” position will disable console access to the CMP (control and maintenance processor). In this position, HP’s equivalent of CNTL-ALT-DEL is disabled. Power cycling becomes the ‘normal’ way to reboot. Jeff Kell adds: Some other systems have a key switch. [Editor’s note: the 9x9s have a key switch on the front which must be turned to “SERVICE” (go figure) to get a CTRL-B prompt.] What is the maximum number of items allowed in an IMAGE detail dataset? Jerry Fochtman and Mike Berkowitz reply: TurboIMAGE raised this limit to 255 ‘fields’ per dataset. However, this is further constrained by TurboIMAGE’s ‘block size’ limit of 2560 half-words, which is the maximum size an IMAGE block can be and contain all the pointers for the entry(ies) along with the block’s entry bit map. Depending upon the size of the field items, one can define < 255 fields in a set, yet not be able to build the set due to the maximum IMAGE block limit. Vikram Kumar adds: Users are also limited by the total number of 1023 fields in the entire database. I know that every connection on a coax (10 base-2) network must be terminated. But can I have an unconnected “T” on a 10 base-2 network segment? The consensus is that 99 percent of the time the network will perform fine with one or more unconnected T’s. In fact, some people even do it on purpose so they can have someplace to plug in test boxes. However, there is also strong sentiment for junking all coax in favor of untwisted pair 10-Base-T. Your mileage may vary. When I execute an MPE IF statement, MPE generates output to $STDLIST showing the result of the IF statement, the commands being skipped, and so on. Is there a way to suppress the output from the IF statement and associated statements like ELSE and ENDIF? Within MPE now, it is a kludge at best of nested command files redirected to $NULL. In 6.0 you will be able to set the HPMSGFENCE variable to suppress the echoing of commands in FALSE command blocks (e.g. if false then ..won’t show the skipped commands). You will also be able to set HPMSGFENCE to suppress the “* EXPRESSION TRUE/FALSE” “* RESUME EXECUTION OF COMMANDS” etc. messages related to IF, ELSE, ELSEIF, ENDIF, WHILE and ENDWHILE. HP says no 5.5-based patch is available for this enhancement. Mark Bixby adds: You could switch now to using the sh.hpbin.sys POSIX if statement, which never generates any extra output. What are the limits for session JCWs and VARs? HP’s Jeff Vance and Scott McClellan reply: Here is a way to find out: `deletevar @ setvar i 0 while true do setvar ![chr((i mod 26)+ord(‘A’))]!i, !i setvar i,i+1 endwhile`The “`setvar ![chr((i mod 26)+ord(‘A’))]!i, !i`” line creates a variable starting with a letter from A-Z with an integer value. On my 5.5 system I can create 10,516 simple variables. Note that this represents a best-case scenario, since I chose short variable names and integer values. The number decreases with long variable names and long (string) values. The example below creates variables with 21-25-byte long names and a 20-byte-long value: `deletevar @ setvar i 0 while true setvar ![rpt(chr((i mod 26)+ord(‘A’)),20)]!i,rpt(‘a’,20) setvar i,i+1 endwhile`and is able to create 8,440 variables. So I take my Windows 95 PC, which connects to our network fine, and plug it into a hub on a newly set up Series 917. I fire up Reflection, try to connect to the IP address, and then wait for up to five minutes to get a login prompt. Once I get connected it seems fine. What kind of things should I consider to get that login faster? Denys Beauchemin and Melvin Rees reply: The problem is one of name resolution. The PC is going through a series of name queries before it tries the name as an IP address. This can be incredibly slow. It could be trying HOST, PROBE, DNS, and finally the IP address. You should verify that you have the HP 3000’s name in the HOSTS file found in the directory of your TCP/IP stack on your PC. Either `\windows` if you use Win95 TCP/IP, or `\reflection` (or the appropriate one) if you use WRQ’s stack. The time-out value is 255 seconds before a different attempt is made at finding the host, hence the long delay. Once the resolution is made, everything will indeed go fine. I’m looking for a means of displaying a warning against unauthorized system access prior to logging on. Our corporate security folks are probably going to mandate such a message be displayed (four to six 80-character lines) on all our HP 3000 systems, and they want to have it displayed prior to the system logon prompt. I know that the ‘logonprompt’ is configurable via SYSGEN, but I don’t think it’s capable of such a message (more than 255 characters). Does any other solution exist? Mike Hornsby replies: On our web site we have a utility called LPRESET that does what you are looking for. It allows you to display a short warning and a prompt. Is there a free (or nearly free) way to send SMTP mail from an HP 3000 to either a Unix Box or NT SMTP gateway? Stan Sieler replies: Yes. Telamon has a mail program that does it free. It works well. Check: ftp://ftp.telamon.com/dist/mail.nm or ftp://ftp.telamon.com/dist/mail.nm.wrq. Note: The Telamon MAIL program must be in a group with PM capability. What does the “:C” or “:S” mean in the VOLUME RESTRICTION column of a store listing? What manual is it found in? Michael Hornsby and Wolfgang Kinscher reply: S is Volume set restriction, C is Volume class restriction. The `NEWCLASS` command in VOLUTIL lets you build subgroups of disks in a volume set; for example, `NEWCLASS BGD:VC1 MEMBER1,MEMBER2.` This can be useful if you want to force specific files to specific members within a volume set. This is then specified with the DEV= parameter of the BUILD command. When you BUILD xyzzy;dev=VC1, the file is then forced to MEMBER1 or MEMBER2 of the BGD volume set. This is all documented in the Volume Management Reference Manual. I have a problem with an account while running in the shell. When I do ‘sh.hpbin.sys -L’ I get: `tput: Unknown terminal “hp2392a”. tabs: Unknown terminal “hp2392a”.`After this, from the shell/ix prompt, I can’t use more, or vi, among others. I get the same kind of error, about unknown terminal. This all works fine from several other accounts, but not from this one. What do I do to make it work? Richard Gambrell replies: Sounds like a permissions problem. As root (oops, I means manager.sys), go into the shell and do: `chmod -R a+rx /usr` This allows all to read and execute everything under /usr, which includes the terminfo files that are needed by vi to control the terminal. Am I using LISTF wrong? When I issue the LISTF,2 command, in the middle of the resulting report I get a line that says “Cannot access file “BAT4IN” due to lock contention. (CIWARN 9169).” HP’s Jeff Vance replies: You must be on Express 3 or 4. LISTF “access” functionality was introduced in Express 3. The short answer is this is a known problem with Express 3. It has been patched (MPEKX18) and is also fixed on 6.0. If you re-issue the LISTF it will likely work, although I know this doesn’t help a bit for LISTFs in jobs. The KX18 patch removes the lock contention problem, fixes a bug in LISTF error handling, and reports accurate EOF for LISTF of files being accessed. How do I point network printer configurations to specific ports on (external) multi-port JetDirect (or equivalent) boxes? Gilles Schipper replies: You need to add the tcp_port_number option, in NPCONFIG, as follows: `(network_address = 128.250.232.40 tcp_port_number = 9100) # for port 1 (network_address = 128.250.232.40 tcp_port_number = 9101) # for port 2 (network_address = 128.250.232.40 tcp_port_number = 9102) # for port 3` Please note that everything on each line after and including the “#” represents a comment. What are the negative values used in FCRANGE in the STORE command to get only database files? For example, to store only databases that start with the letters ONRL@ you would say: `:STORE ONRL@;SHOW;FCRANGE=-n/-m;DIRECTORY;`.........etc. [Ed. note: Several people posted, and many more, myself included, probably readily agreed that the answer is -401 and -400. Stan Sieler set us all straight:] Someone just posted -400 & -401, which is the old answer – but only for some databases. Here’s the list of -4xx filecodes I have: ! # Mnemonic ! Usage First seen? !–- ––– ! –––––––––––––&# 150;– –––––- ! HP defined privileged filecodes (SS mnemonics)... -400, “iroot “, ! image root -401, “idset “, ! image dataset -402, “irdba “, ! image remote database access -403, “idump “, ! image i-file (and j-file) -404, “ilr “, ! image intrinsic level recovery -405, “iglob “, ! TurboIMAGE/XL control block file -406, “idrb0 “, ! TurboIMAGE/XL DRB’00 file -407, “igtx “, ! TurboIMAGE/XL GTX file -408, “iccf “, ! TurboIMAGE/XL Jumbo chunk control file -409, “icdf “, ! TurboIMAGE/XL Jumbo chunk data file -410, “itpir “, ! TurboIMAGE/XL TPI “root” file (was: omni1) -411, “itpid “, ! TurboIMAGE/XL TPU “data” file (was: omni2) -412, “iidx “, ! TurboIMAGE/iX ..idx file -420, “irecv “, ! DBRECOV restart file -421, “chgcf “, ! DBCHANGE control file ! SS 940426... (from Goetz Neumann, HP Germany) -430, “tc “, ! TC file (IMAGE root name + “TC”) links root -> DBE -431, “atc “, ! ATC file (links a DBE -> IMAGE roots) -450, “dausf “, ! DAUSF is the NETUsers config info (SC 940930) -451, “daalf “, ! ?? created by NETUTIL (SC 941001) -491, “dbe “, ! DBCORE database environment (“DBDB” in first 4 bytes) ! ALLBASE_DBCON_FILECODE -493, “ptpt “, ! DBCORE Chunkset (“PTPT” in first 4 bytes) -497, “lflf “, ! ALLBASE (DBCORE) Logfile (“LFLF” in first 4 bytes)Is there a manufacturer of a JetDirect-compatible print server (i.e., one that will work with MPE/iX 5.5’s network printing) that has one or more serial ports? As far as I know, HP does not make a JetDirect Ex box with a serial port. I would like to eliminate a number of remote DTCs (and the bridging they require) whose only function in life these days is driving serial printers. Life’s tough. Either use a JetDirect Ex with parallel-to-serial converters, or use a third-party print server with serial ports and third-party software to drive it.