Q: When upgrading an old system from 5.0 to 6.5 are there any mandatory versions in between one must upgrade to first before getting to 6.5?

A: Tom Hula replied:

According to the System Software Maintenance Manual, you would have to upgrade from 5.0 to 5.5 or 6.0 first. Then you can upgrade to 6.5. There is no way to get directly to 6.5 from 5.0.

A: John Burke replied:

The only way you can successfully print to your Okidata printers over the network from the HP e3000 is with one of several 3rd-party products. HP has steadfastly refused to enhance its network print spooler, deferring to the 3rd-party vendors.

Q: I'm having problems using Reflection 7.0 to upload files to an e3000 (6.5pp2) across the Internet. I have tried both telnet and VT without success. The transfer just seems to hang.

A: Joe Silagi replied:

The problem is likely a firewall interfering with Fast File Transfer. Reflection's Fast File Transfer (FFT) uses a separate TCP connection. On the pc the port FFT uses is 30767. On the host side the port number varies but is in the range 30000-40000. Firewalls will need to be configured to allow access to these ports. If the FFT negotiation fails (default behavior) Reflection will use the session port for the file transfer. In the File Transfer Setup you can configure Reflection to negotiate (default), always or never use FFT. You may need to disable FFT.

Q: I need to verify within a STREAM that a file does not exist. If it does exist then abort, else finish the job. I am trying to make sure a file is not overwritten.

A: John Burke replied:

You want the finfo CI evaluator function. Just type HELP FINFO at a CI prompt for all the gory details; for example,

if not finfo("fname","exists") then
...

Q: Today I got a new PC with Windows 2000. My password protected Samba Shares do not work any longer. The other unprotected shares work like a charm. Any hint to solve this problem?

A: Michael Gueterman replied:

Make sure you have applied the Win NT 4.0 registry patch (which works just fine on Windows 2000). Also, if you are using a user.map file, you may need to make some alterations to it. I've included that information, which was provided by Greg Stigers, in with some other Win2000 notes in a document entitled "Windows 2000 and Samba/iX Issues" available at the sambaix.com web site. You can get to it from the Document Library section of the site.

Q: I've checked the archives of HP3000-L for the "scrollable pick list" enhancement to VPLUS (which I was sure I had seen something about here) and could only find a post from 1996. Has this enhancement been done or not?

A: Mark Undrill, Nick Demos, Lars Appel and Michael Berkowitz replied:

It has not been done. After starting, it HP decided not to do it and rely on third partys.

Q: What is the upper limit of LDEV numbers?

A: Stan Sieler and Christian Lheureux replied:

Depends upon the version of MPE/iX:

1.1 .. 3.1 MAX_LDEV = 999
4.0 .. 4.5 MAX_LDEV = 2330
5.0 .. 7.0 MAX_LDEV = 4679

Stan Sieler added: I wouldn't be surprised to find that some code doesn't like disk drive LDEVs larger than 999.

Q: Can I use the ALLOW command without rebooting the machine?

A: Ron Horner, Andres Ogayar and Ted Ashton replied:

You can issue the ALLOW command anytime you want. You can only issue the command at the console. You can 'get' the console at your current session (you must have SM capabilities), with the command 'CONSOLE <nn>'. After that, issue the ALLOW command and return the console to its origin 'CONSOLE 20'.

Jeff Vance added: The ALLOW command itself can be allowed.

Editor's note: Not only does an ALLOW not survive a reboot (so the original questioner was in for a lot of frustration), it is only valid for the duration of the job or session. A persistent ALLOW facility was on the most recent SIB but did not make the top ten.

Q: Is there a way to empty Apache and Samba log files? Mine go back quite a few months and are quite big. Can one 'PURGE' these files and expect the deamon/listener to rebuild them? Is there a POSIX way to empty them?

A: Mark Bixby replied:

Apache has a rotatelogs utility that will automatically create a new logfile after some fixed time interval:

http://httpd.apache.org/docs/programs/rotatelogs.html

MPE :PURGE should complain about log files being in use. But if you do a POSIX rm, the log file will be purged, but what happens next is probably not what you expect. The log file doesn't REALLY get purged until the last accessor has closed it, and so the network server will continue to write entries into the now "hidden" log file, consuming additional disk space. Once you terminate the network server that is writing to the log file, the file will finally be removed from disk. So it's cleaner to clean up your log files when the relevant network server isn't running.

Many POSIX network servers however will allow you to do something like this while the server is still running:

mv logfile logfile.old (rename the active logfile)
kill -HUP nnn (send the HUP signal to pid nnn to cause a server restart)
optionally process the now closed logfile.old
rm logfile.old (purge when done)

Apache works this way.

Lars Appel added:

While I've never used it myself, SAMBA seems to have a "max log size" directive in smb.conf that causes a log switch to log.old or similar and a fresh one started.

Q: I've modified the configuration for our tape drive in sysgen to mode=autoreply, however, when I do a backup, I'm still getting a reply request. Is a system reboot is required?

A: John Burke replied:

In a word, yes. A START NORECOVERY is required after re-configuring a tape drive to AUTOREPLY.

Q: I have a system on MPE/iX 6.5 PP2 with a serially connected (via DTC) 2563B (recently moved from one floor to another). It is now not reporting "out of paper" or "paper jam" to the 3k. I am told that this printer used to stop (Flashing 11) and wait for more paper, but now it just happily prints everything whether there is paper or not. Did I miss anything on changes to the spooler?

A: John Burke, Andres Ogayar, Doug Werth and Barry Durand all replied:

The "paper out" sensor is probably either faulty or dirty.

Q: I have forgotten how to set dip switches on a Jamaica box to a particular SCSI address. Could someone refresh my memory?

A: John Burke replied:

4 bit binary. Up is one and down is zero.

dddd == 0000 == 0
uddd == 1000 == 8
uuuu == 1111 == 15
etc.

You'll probably need a flashlight, a pointed tool and, also, young eyes because the switches are inhumanly small.

Ken Sletton added:

While you're doing this with the small pointed tool and reading glasses (alternative to 20-something eyes), I'll add my $0.02 follow-on suggestion: "Rock" each switch back and forth a couple times; and try to note if all switches have about the same "tactile feedback". Because they are so small and even good ones have low resistance to movement it can be a little hard to tell, but this let me catch a potential address conflict problem. We have two HASS enclosures. On one of them, one of the DIP switches was so "loose" you could almost blow on it and cause it to "fall" to the zero (down) position.

My solution? In lieu of major repair (if I understand correctly HASS DIP switches are "soldered" to the back plane - or so I was told), "embed" the defective DIP switch in one drop of RTV silicone. Of course that makes it a bit harder to change again if we should ever need to do that, but in our case, odds were low enough I went for the quick and reliable solution.

Q: I've got a question about the CI function 'finfo'. There are two options available that reference blocksize. One is called 'blksize' (#15). The other is 'byteblksize' (#31). Anyone know what the difference is between the two?

A: John Clogg replied:

The short answer is that 'blksize' returns words and 'byteblksize' returns bytes.

Q: Is there a way to cause a specific user to automatically logoff the HP 3000 after a specified amount of time in which they are inactive? Say 15 minutes?

A: Ken Sletton replied:

SETVAR HPTIMEOUT 15

Jeff Kell cautioned:

Note that HPTIMEOUT only applies at a colon prompt :-(

Q: How can I configure a Default Gateway?

A: Fred Metcalf replied:

On screen #158 in NMMGR: Neighbor Gateway Reachable Networks (Path: NETXPORT.NI.SQUHNET.INTERNET.HPRS )

Press the 'Help' key (F7) and read this:

IP network address

List the IP addresses of all the networks that you will be able to reach through the gateway you are configuring. You also use this field to indicate whether or not the gateway is to serve as a default gateway by entering an at sign (@) to specify that it is a default gateway.

Q: Is there a tool which given an IP address will return the mac address of the device responding to that address?

A: Karl Hancock replied:

You can look up the currently known ip addresses and mac addresses in the route tables using nettool.net.sys.

:nettool.net
>>>nameaddr
>>>mappings
>>>mapping

and then find the IP address of interest with its corresponding mac address in the table.

Gavin Scott elaborated:

This is called the arp (address resolution protocol) protocol. It will only work for devices on the same 802.[x] or Ethernet LAN segment however, since MAC addresses never pass across network boundaries.

For devices on your own segment that your 3000 is talking to, you can use the NETTOOL program (as someone gave details for) to look at the arp cache of mappings from IP -> MAC.

Q: I can't get to jazz from work. I suspect it may be due to the way our network is configured. The numeric IP for jazz is 192.6.38.5 and all of our internal IPs are 192.6.*.* . Is there a way for me to get to jazz from work?

Mark Bixby replied:

The jazz subnet is 192.6.38.*. Non-HP networks should not be using these addresses if you want to be able to access that HP subnet.

Internal networks should not be using IP addresses which conflict with the outside world. A special block of "private" addresses was created to prevent this conflict problem. It sounds like you need to consider renumbering that part of your internal network, and you should consider using the address ranges specified in RFC1918:

http://www.faqs.org/rfcs/rfc1918.html

Q: I have an HP e3000 with a Session Limit of 60 and, in SYSGEN/MISC, a Session MAXLIMIT of 100. However, if I have 20 sessions running and the 21st session tries to log on I get:

USER LIMIT REACHED; LOGON REFUSED. (JS 903)

Does this mean I've reached my MPE license limit?

A: Gavin Scott replied:

I believe so, yes. You can look at the HPUSERLIMIT variable to see what the "license" limit is for the machine.

Note that there are some things which count as "sessions" but which do not count as "users" against the license limit, so generally you want to have your session :LIMIT larger than the HPUSERLIMIT value, otherwise you may artificially limit the number of useful connections you can have.