Q: I'm a new sys admin. I started this job
in July, just before the new N-Class came to replace the 987 as a
development/production shadow machine and we upgraded the production 969 to
7.0. Now we've got this perfectly good 987 on our hands, and my boss is
wondering if we can move it to our out-of-state site and use it for disaster
recovery. But we would need to upgrade it to 7.0 in order to shadow (Shareplex)
our production machine to it, right? Can we even shadow to two different
machines in the first place?
A: John Clogg, Ron Horner and John Burke
replied:
7.0 will not boot on any 9x7-series system.
There is, however, no technical reason for it not working. HP marketing simply
decided to try to goose up sales of new systems by forcing the obsolescence of
the 9x7s. [As of HPWorld 2002, HP is still insisting it will not change this
policy.] 6.5 is as far as you can go on a 9x7.
As for Netbase/Shareplex shadowing to
multiple machines, absolutely it can do that. Of course you'd need the license
for the extra machine. Not that the shadow machine can be on 6.5 even while the
master machine is on 7.0.
Q: We moved an Allbase dbe/database from one
system to another. Now when the users attempt to connect, they get:
START DBE disallowed: DBEnvironment points
to a wrong log file. (DBERR 3106)
What is the problem?
A: Bill Cadier and Russ Rino replied:
The log file wasn't restored with the
database/dbe so you'll need to do a START DBE NEWLOG, for example:
:ISQL
isql=> start dbe 'DBENAME' multi newlog
log dbefile DBEdblog03 with pages=1000, name='DBElog03';
Q: All of a sudden I've got a problem with
Apache on my 6.5 PP3 system. The apache version is 1.3.9 and, of course,
nothing has changed. Anyhow, if you try to connect to this machine's home page
it won't load and eventually times out. I'm seeing the following in the
error_log:
[Sat Aug 31 09:39:53 2002] [error]
(11)Resource busy, try again: fork: Unable to fork new process
[Sat Aug 31 09:40:04 2002] [error]
(11)Resource busy, try again: fork: Unable to fork new process
over and over and over. The machine was
rebooted early this morning. All other network services (like telnet) appear to
be OK. There's nothing notable in the latest nmlg file. The machine is
bascially idle (I'm the only person logged on other than the operator). Also,
the httpd server seems to not respond to the usual 'kill' command but it can be
aborted. Any clues?
A: Gilles Schipper replied:
I seem to vaguely recall issues related to
Apache, forking, and moderately fragmented disks. How fragmented are your
disks?
[Editor's note: This turned out to be the
problem.]
Q: I have a question regarding the build
command. Last night I had a job fail because the build for the file failed. The
command was failing, complaining of no disc space. This was the command.
BUILD NEWFILE;DISC=524288,1,1;REC=-384,21,F,ASCII
What puzzled me was that discfree showed
more than 40 million sectors free. Then when I removed the ,1,1 from the DISC
parameter, the build worked. I am just curious as to why. I know the ,1,1 has
to do with extents, but I never learned the reasoning for them, or what they
really mean.
A: John Burke, Mike Church and Tom Emerson
replied:
You were trying to allocate all the space in
one extent (that's what the "1,1" means); i.e. one contiguous block
of storage. Undoubtedly, you disk is fragmented sufficiently that this was
impossible.
Q: Does anybody know the file size limit for
a RIO file? My system (6.5) gives me the FSERR 106 message when trying to build
the file. I tried specifying the number of extents and got the same result.
:BUILD
LRR01003.group.account;REC=-187,,F,ASCII;RIO;DISC=550000
EXTENT SIZE EXCEEDS MAXIMUM (FSERR 106)
:calc 550000*187
102850000, $6215DD0, %610256720
A: Michael Berkowitz noted:
RIO files are one of the few file types that
are not in native mode, which means they are limites to 1/2 GB.
John Burke replied:
Actually, this a somewhat of a misleading
error message. You have fragmentation "issues". Try
"DISC=550000,32,32" (the maximum number of extents is 32) and see if
that does not work. If it does not, then you REALLY have fragmentation problems
(assuming you have adequate total disc space) that you need to address either
through a total reload (ugh) or one of several software utilities that will do
defragmentation.
Q: We got two DDS-3 tape drives in August
and ordered 125-meter tapes. We have been using the 120-meter tapes we had for
the old tape drives until the new ones came.
The new tapes came today. The vendor sent
150-meter tapes that say DDS-4, instead of the DDS-3 tapes that we ordered. Can
they be used in the DDS-3 drive anyway?
A: John Burke and Warren Bunnell replied:
No, in fact they will not even load.
Denys Beuchemin added:
You state that you have a DDS-3 drive and
that you have been feeding it 120-meter tapes. I am sure by now you realize
that you have been using the DDS-3 drive as a DDS-2, same speed and recording
capacity (1MB/second for 4GB). In order
to obtain the real throughput of the DDS-3 device and the recording capacity
(2MB/second for 12GB), you must use DDS-3 125-meter tapes.
Now for the 150 meter tapes. Indeed, you
cannot use them in the DDS-3 drive. The length really has nothing to do with
it; rather the 150-meter tape is thinner than the 125/120/90/60 meter tapes and
is subject to stretching and breakage in a non-DDS-4 device. DDS-4: 5.6
microns, DDS-3: 6.8 microns.
Q: I now have the newest supported version
of Samba running. I changed my Samba startup jobs to match the supplied jobs.
The problem is, how do I stop the jobs with resorting to abortjob?
The relevant line from the new job (for nmb)
is:
!xeq nmbd207.samba.sys "-D -p 137"
A: Lars Appel replied:
Try using the pid files
/SAMBA/SMB207/var/locks/?mbd.pid for kill. Samba 2.x creates those .pid files
automatically during startup.
Q: Does anyone know why y system clock would
be running slow? We have a 979-200 on MPE/iX 6.0 PP2. It loses about 1 minute
per day.
A: Bill Caudier replied:
If your system does not have MPEKXY0
installed you should consider installing it. This patch has been superseded a
few times so if you have MPELX38, MPELXG5, MPELXK9 or MPEKXR0 installed then
you would have MPEKXY0 too. None of these patches was on 6.0 power patch 2.
If you already have one of these patches
then the problem could be hardware related so it might be worth contacting your
hardware support people to check that out.
Bob J. added:
Another possibility was addressed by a
firmware update. The text from the current cpu firmware (41.33) update
PF_CMHK4133 mentions:
- System clock (software maintained) loses
time. The time loss occurs randomly and may result in large losses over a
relatively short time period. Occurrences of the above problem have only been
reported against the HP3000 979KS/x00 (Mohawk) systems. Software applications that perform frequent
calling of a PDC routine, PDC_CHASSIS, affect the amount of time lost by the
system clock. Your hardware support company should be happy to update for you.
Tongue firmly in cheek, Wirt atmar noted:
My first guess would be relativistic time
dilation effects as viewed by an observer at a distance due to the fact that
you're now migrating off of the HP3000 at an ever accelerating rate. My second
guess, although it's less likely, would be that your machine has found out that
it's about ready to be abandoned and is so depressed that it simply can no
longer work at normal speed. We've certainly kept this information from our
HP3000's. There's just no reason that they need to know this kind of thing at
the moment.
And, in the same vein, Bernie Sherrard
added:
Look at the bright side. At a loss of one
minute per day, you won't get to 12/31/2006, until 2AM on 1/2/2007. So, you
will get 26 hours of support beyond everyone else.
Q: I am trying to create a new database with
two detail datasets and 5 automatic masters. DBSCHEMA is reporting the
following error:
***** ERROR ***** MORE THAN ONE PRIMARY MASTER
What is my problem and how do I fix it?
A: Denys Beauchemin, Michael Berkowitz and
Jerry Fochtman replied:
It means exactly what it says. You have more
than one item defined as the primary chain in a detail data set. This is done
by putting a "!" in from of the master dataset that links to a detail
index.
Q: Some time ago, I recall seeing a
reference to commands that would remount a DDS cartridge, without the operator
having to physically take the tape out and put it back into the drive. I
searched the Archives, but did not find what I was looking for. Can someone
please supply the script, or point me to the correct commands?
A: Paul Courry replied:
DEVCTRL.MPEXL.TELESUP
Just type in the name without any parameters
and it will squirt back the instructions for use.
Stan Sieler added:
In addition to DEVCTRL, you might want to
check out our free ONLINE program at
http://www.allegro.com/software/hp3000/allegro.html#ONLINE
Q: We are trying to access a database on our
old system using QUERY and it is asking for a password. I have done a listf ,-3
on the database, but there is no lockword listed (which I assumed would be the
password). Can anyone tell me where to find the password assigned to a
database?
A: John Burke replied:
Assuming you do not have access to the
original schema and you want to know what the password is, not just access the
database, then sign on as the creator in the group with the database, run
DBUTIL.PUB.SYS and issue the command:
SHOW databasename PASSWORDS
Mike Church and Joseph Dolliver added:
If you just want to access the database, log
on to the system as the database creator and, when asked for password, put in a
";" semi-colon and hit return.
Q: We have a new HP laser printer that will
need to be accessed via our network and the 3000. We don't have any ldevs
available. What process do I need to follow to set up a new one? Do I set it up
in SYSGEN and NMMGR? Any suggestions would be greatly appreciated!
A: John Burke and Gary Jackson replied:
You do not need either SYSGEN or NMMGR. You
can use IOCONFIG and dynamically add network printers. Also, as far as an LDEV,
you will be able to choose any number not currently used. This was all first
documented in the MPE/iX Communicator 5.5 (see
http://docs.hp.com/mpeix/onlinedocs/communicator/ix55/main.html for details).
Q: Does anyone know of a command for
aborting all user ssessions? We're trying to develop a controlled way of
ensuring no-one is logged on prior to system backup.
John Clogg replied:
The LOGOFF command will do it, but that's kind
of a big hammer. Note that LOGOFF will
also abort the session that issues the command, unless you specifically exclude
it. It also prevents any new jobs or sessions from
logging on, so it must be followed by a
LOGON command. Both commands can only be issued from the console, I believe.
Jeff Vance added:
The ABORTJOB script (and/or UDC) on Jazz at:
http://jazz.external.hp.com/src/scripts/index.html could work for you. If you
download the script and name it ABORTJ, you could do the following: :abortj @
This will abort all jobs and sessions on your system, except for
"you".
Q: We recently installed an N-class machine
and have an interesting problem. We do not get any messages on the system
console. No tape mounts, no TELLOPs, nothing.
I have checked to make sure the console is,
in fact, assigned to the console (LDEV 20). So that appears to not be the
problem. And yes, it is logged into the system as OPERATOR.SYS. The system
otherwise seems to be operating just fine. You can answer console messages, but
the only way to see them is with the RECALL command.
Does anyone have any ideas? I'm really scratching my head on this one.
A: Elizabeth Campbell and Jon Backus replied:
It sounds like Easytime. If anybody has even
run Easytime interactively it will cause this to happen. The batch job doesn't
even have to be running for the console messages to be diverted.
Q: Building a 989 from scratch, using 6.5. I
did the system load tape. Ran Autoinst, number 3 for FOS, SUBSYS, and PP3. It
seemed to restore the FOS correctly, SUBSYS went well (about 34 jobs), but when
it got to the part where it backs up the SL, XL, NL, it aborted because it
could not find the file AUTOGEN.INSTALL. It does not exist. Any suggestions,
besides call an expert?
A: Ron Horner and Jeff Kell (who are both
experts) replied:
:RESTORE *pp3tape;@.install.sys
Q: Does anyone know if a HP LaserJet IIIP is
networkable?
A: Wirt Atmar replied:
Although I'm tempted to write: "As in
hooking a chain to it, as in a boat anchor?", I won't write that.
The answer I truly suspect you want is: any
HP PCL-speaking printer that has a parallel port on it can be networked into an
HP3000 through the use of an external JetDirect box. If the LaserJet IIIP has a
parallel port on its back,
it should be quite easy to do.