doesn't *solve* the problem, but neither does
the word war we've waged. Work the problem.
Folks, we need to think carefully. Nothing wrong with argument.
Just be constructive. Doesn't matter whether it's Alan Cox
or Alan Altmark, each man has gotta eat. Argue productively.
--
Rick Troth, BMC
that
a module built for 2.4.9 won't work with 2.4.9-4GB? What patches
are these that warrant changing the label??
--
Rick Troth, BMC Software, Inc.
2101 City West Blvd., Houston, Texas, USA, 77042 1-800-841-2031
When installing Redhat, interrogatories are used to form the parameters of
the network. For example, the the IP address to be used. This address is then
stored in a file /etc/sysconfig/network-scripts/ifcfg-ctc0 (if using CTCs).
Where is/are the DNS server address, the Gateway address, and
On Thu, 17 Jan 2002 [EMAIL PROTECTED] wrote:
Ah - so that begs the question - how do you have a root shell running on
the console without signing on first?
I had one in my hints and tips talk from this past summer.
It is embedded below. Change your inittab to run this
instead of 'sulogin'
On Thu, 17 Jan 2002, Rob van der Heij wrote:
I see that this logon through the console still does not show
up in 'who' and I suspect that is related to this message?
Some program (probably getty (one of mingetty, agetty, etc))
must maintain the utmp and wtmp files. One lists who IS logged
On Tue, 22 Jan 2002, Coffin Michael C wrote:
EXACTLY! Forcing service contracts down the throats of shops that are
evaluating ...
SuSE does not seem to get!
What is interesting to note is that this S/390 strategy
appears to be the opposite of SuSE's INTeL Linux strategy,
where they (like
...
where it is really a misnomer: the JVM is an interpreter.
z/VM and VMware are true virtualizations.
What is different about z/VM and the JVM.
The JVM exists in silicon form as well as on paper.
Hercules is not virtulization.
Yet the instruction set implemented by Hercules exists in
-) Linux on zSeries does not have the look and feel of Plain Linux
I wonder how much of this is driven by the line-mode characteristic
of the console (specifically HWC and emulated 3215, but also
effectually when 3270 driver is employed). In fact, it is
not intractable to get
On Thu, 7 Feb 2002, John Summerfield wrote:
Programs that have builtin ideas about where their components are will
fail if their expectations are not met; gcc, perl and exmh are
candidates (I suspect they're okay) as is dhcpd which I suspect will
fail.
In the Hints and Tips talk I gave at
...
to know if there is a way for all three of them to share the same disk
space (I mean to share a partition between all of them). ...
We do this all the time.
The easiest candidates for sharing are /opt and /usr.
By default, they get put into the root filesystem,
so you need to give them
http://www.corestore.org/intro.html
THIS could be PROMISING.
When we all run out of work, we can swipe Mike's
gas turbine to power our surplus 9672s and 3090s and what not.
Ross closed the loop, so I'm reopening it. evil grin
On Wed, 6 Feb 2002, Jim Wenzlaff wrote:
I have read about the patch on Developerworks, but it is still somewhat
unclear to me where to start. Can the patch be installed on the Redhat or
SuSE distributions?
I would expect the patch to
On Thu, 14 Feb 2002, John Summerfield wrote:
Still there will be configuration choices that may impact how well the
binaries fit in.
But that should not make a kernel built from the public generic
source tree not work on any given distro. (As you say further on.)
If Dingo Linux has ext3
root@t2linux:/home/khaam lpc start npf0010
npf0010:
printing enabled
lpc: connect: Connection refused
couldn't start daemon
Is the server up on the remote end of npf0010?
We need a Lingua Franca for hypervisors.
Consider the command
hcp attach F200-F202 mylinux
Makes perfect sense, though the handle is a zSeries I/O range.
What would that mean to INTeL? Might look more like
Did I miss something? That address range is valid on
Is there any problem with top under Linux/390 (s390/s390x)? It's performance
is hideous with anything like 100 processes in the system. top usually
becomes the biggest CPU user. I don't see the same on non-S390 platforms
with equivalent numbers of processes in the system.
With or without VM?
anyone know of an ftp client that allows to:
1. tell in a profile to transfer .html .txt (text files) as ascii
AND all others as binary (defaults binary)
Sounds like any reasonably designed web browser.
2. transfer an entire directory tree containing mixed file types
Dunno
Can anybody advise me on defining virtual CPUs for use by
a 2.2.16 level Linux kernel? Does it buy me anything to give
a Linux image more than one (virtual) CPU, or should I just
give each image one and let CP do the multiprocessing? TIA.
I don't think it buys you anything
other than to
On Tue, 26 Feb 2002, Jon R. Doyle wrote:
/etc/rc.d/init.d ? Oh, right that is the runlevels on RH. Solaris, SuSE
blah blah use /etc/init.d think that is system 5 or LSB some such
standard, not sure why RH adds the other layer.
/etc/rc.d/init.d should be (as in if it is not, then make it
In my case I was copying the /var tree and it left out /var/run/printer
and /var/run/.nscd_socket which are both symbolic links to elsewhere.
You sure they were sym-links and not socket files?
Sym-link should have an l in the far left of 'ls -l' output.
A named socket would have an s in the
What
does the .bz2 extension signify?
BZip (or BZip2). It is replacing GZip in some areas.
In fact, some FTP sites are going so far as to offer both
.gz and .bz (or .bz2) compressed copies of whatever
not unlike they did some years ago with both .Z and .gz.
Meanwhile ... At the Oscars,
Robert Redford proclaims freedom. [sigh]
When I said it should be retooled
what I meant was that for some HW platforms (like zSeries)
there is a hardware concept of a processor serial number.
On such platforms, 'hostid' should report that value
or should report something derived from it
and not something derived from IP address or
I was wrong.
I misunderstood the nature of gethostid()
and therefore misunderstood the relationship of 'hostid' to it.
Thanks to those who have indulged me in re-education.
Trying to download some patches today.
The little widget that forces me through the licence agreement
seems to be broken. It's getting a file not found every time.
This is with Netscrape on PC Linux,
Internet Exploder on NT and Netscrape on NT.
Would anyone have any recommendations?
LISTSERV.
See http://www.lsoft.com/
Have SUSE 7.0 lpar, and will have a RH 7.2 soon, ...
Our email servers are NT.
LISTSERV runs on NT and on (PC) Linux, and several others.
Dunno the status of their S/390 Linux port.
Well,
if you don't want to spend any money,
you might could get away with ye olde SENDMAIL aliases.
An alias can be defined by a file, and the file amounts to a list.
It is trivial to crunch a LISTSERV-format list
into an /etc/alias format list:
cat LISTSERV-list grep
...
but make gets error in system.h with error messages:
/usr/include/asm/system.h: In function '__xchg':
/usr/include/asm/system.h:67: '__u32 undeclared
/usr/include/asm/system.h:67: parse error before 'ptr'
/usr/include/asm/system.h:90: parse error before 'ptr'
make[2]: ***
did not spot -ansi flag anywhere in compiler flags.
You might be tripping over something different.
What I specifically thought it might be is in linux/types.h
there is a section that reads something like
#if defined(__GNUC__) !defined(__STRICT_ANSI__)
typedef __u64
On Tue, 16 Apr 2002, Mark Earnest wrote:
On Wed, 17 Apr 2002, Alan Cox wrote:
... making everyone think of balding old farts who speak only JCL 8)
Hey, not fair! I happen to know one MVS systems programmer who is a
balding 23 year old fart :)
You telling us that MVS causes premature hair
I hope this is not too vague, but is there an established way to send a
message from Linux S/390 that will show up in an OS/390 LPAR's logs?
The question makes me think of ye olde NJE networking
with its ability to write to a remote console. Write to logs?
I don't know, not knowing MVS well
Gregg ...
I find that you can generally upgrade your kernel
without too much harm. Suggest that you get the 2.4.18
and use it on your otherwise 2.4.lower based system.
This is INTeL? Or is this S/390 Slackware?
I said (to John):
S/390 does also have FBA (fixed block architecture) DASD devices.
Sadly, even in the Linux world these are not widely known or used.
I maintain that they should be employed heavily and heartily!
The benefits are numerous.
On Thu, 18 Apr 2002, Jay Maynard wrote:
What Dave Boyes said about SYSLOG is a good point. Also:
If we ever get VM here,
is there a linux - VM - OS/390 method that's better?
Sure. Presuming you have an NJE link between VM and OS/390,
something like this:
hcp smsg rscs msg mvshost mvsopcon something-to-say
Mike ...
If RedBooks are copyrighted,
please just double check that publishing this chapter is
within the bounds of that copyright.
Has anyone heard of any IBM plans to provide non-ECKD DASD
support for zOS?
We can't go on emulating 33X0 volumes much longer.
What Harry said.
While I agree that MVS should grok FBA, but
the problem is not that we have and continue to have CKD.
The problem is that FBA is so poorly
Bad news.
This is particularly bad news to those of us who use VNC.
UK version of PARC. The Cambridge lab is where VNC was created.
Too bad some other techno giant hasn't seen fit to fund this lab.
-- Forwarded message --
Date: Wed, 24 Apr 2002 14:48:05 -0500
From: Hal
Please do NOT peer the newsgroup (if any)
with this mailing list.
I agree with Jay's analysis/description, but ...
Rich asked:
Pardon my ignorance in this matter, but what's the point? We already have
a
great group here and I would personally prefer a single point of
discussion.
Do newsgroups work any way like mailing lists such that messages are mass
On Mon, 6 May 2002, Coffin Michael C wrote:
You're going to need to SEND those Linux commands in lower case (by default
they will be upper cased). Use EXECIO to pass the command to CP ...
I have to recommend Diag(08,'SEND'...) or Diagrc(08,'SEND'...) instead.
This is probably the Pipelines
I ran across secuser and send while reading this weekend and have tried to
create a rexx exec that I could use to automate the shutdown of a linux
image but I can't seem to get the login to work. Has anyone an example of
doing something like this under z/vm that they would be willing to
...
said, you then need to get concerned with the default PIPE separator stage
...
You could probably deal with that issue like this:
/* REXX */
Parse Arg cline
'PIPE VAR CLINE | SPEC /SEND SUSE0001 / 1 1-* N | CP | CONSOLE'
Exit rc
I guess by even
On Fri, 3 May 2002, Rich Blair wrote:
1. Give each development group their own Linux. ...
2. We (IBM syprogs) micromanage each Linux - ...
I'm very much in the #1 camp.
In fact, I would not stop at the group level
but would go so far as to give every *user* their own Linux.
(This does
[leaving SAMBA on CMS in better hands to respond]
Is there any way for CMS and/or Linux users to access data on each
other's filesystems directly (as files, not using the raw I/O PIPE
stages)?
There's the CMS FS package,
which is both a utility (to read CMS files directly from Linux)
and a
John, I truly do not think you need to worry so much about UUENCODE.
I suggest you remain skeptical, but not about UUE. It is only
one example, and is much less problematic than others.
I think I used CP500 on OS/2.
I think CP500 is an EBCDIC codepage, not an ASCII codepage.
(Loose use
I believe that the gentleman mentioned that he has no VM available,
so no PIPE solution to this problem..:-(
Thanks for noticing, Dave, but check again.
I specifically referenced TSO Pipelines, not CMS Pipelines.
Aside from the performance issues already discussed
(and Rudy, I trust you were copied on those responses
or are signed up to the list so that you would see them),
running SETI@home on zSeries is a mistake where zSeries Linux
is running in a virtual machine. SETI@home attempts to reclaim
CPU
As long as it simulates 3390 and ECKD, ...
And if it could simulate 3370 (or 9336, or ...) that would be better.
Alan said:
Maybe someday VM will itself be able to use SCSI instead of ECKD,
but that day has not yet arrived. ...
I would like to think that this new option
would present SCSI as FBA, which VM already supports perfectly.
But ... from the sound of this discussion, there may be
Willem always has a great response
and noted to me off-list that the kernel source appears to
present this new FCP thing withOUT any translation. So it's
better than FBA. Fond as I am of FBA for zSeries and esp Linux/390,
I must admit that eliminating emulation is a Good Thing [tm].
this one had LINUX inside the fish!
Perhaps a messianic view of Linux?
Mike ...
Best to compare the problem account and host
to a working user@host pair. Do you have SSH working anywhere?
So the first thing I would recommend is to back away from
this particular one and see if you can get that SSH client
working to some other Linux (or UNIX) host with any account
Probably not what you want to hear, but:
- Looking in the termcap or terminfo databases. The termcap and
terminfo sections for 3270 are quite incomplete. As a result,
programs like /bin/clear don't function properly.
Presumes that the console driver would
BY THE WAY ...
I maintained for years, and still do,
that a 3270 can be used effectively from carefully written
ASCII application programs. The big requirement is not
3270 data streams nor EBCDIC, it's block mode toleration.
Tools like YaST are awfully close to being able to do this.
The
The other question is much more interesting. Is there enough pent
up demand
for then triple redundancy and all the other fault tolerance in
the mainframe
world to create a mass market for it.
Yet again, Z SEEries phoned home.
Over the weekend we had one of those outages where
before
The UTS Global 3270 line-mode driver was designed to handle many
commonly encountered ESC sequences too, for example those generated by
the ls command in its full-color glory; ...
It is conceivable that applications like 'yast'
could be modified so that what they see as a block mode ANSI
To:[EMAIL PROTECTED]
Good. Send the command to the robot. That is correct.
subscribe linux-390 Bob Tegethoff - at Pepco
Good. The syntax is correct too.
Question:
Was the note sent as plain text?
If it was, for example, HTML, that will throw LISTSERV off.
What David Rock said.
And I don't think VMware is at all to blame.
I *suspect* that virtualization of Pentium is hard to do.
zSeries has more than three decades of hardware and software
adjusting for each other to do virtualization very efficiently.
I use VMware all the time. Not continuously
We need the STM at the start of kernel execution,
immediately after the bootstrap finishes.
It's a pity this has never been implemented!!
Delighted to hear this news.
I used 'ned' on UTS some decade ago and more. Nice!
Will y'all be porting NED to PC Linux?
As I recall, running 'ned' on a VT100 worked reasonably well.
(I never did; I heard from others; I always used real 3270s.)
So while a PC would not often have a real
http://www.theregister.co.uk/content/4/26230.html
I wonder what the odds of an MS-Linux distribution are?
My compliments to IBM for not producing an IBM Linux distro.
I fear that Microsoft will be much more disruptive to Linux.
IBM stood to gain more (still does, I say) from
Is anyone running a combination of Java, Oracle, Weblogic, on Linux/390?
If you are, would you be willing to talk with someone else
running a similar configuration?
Thanks!
Oh, you can't do that. It's not supported; you have to use NFS for that.
The problem is not with VM or Linux,
but with software management tools and really more with
how the packages are built by the package creators.
RPM, for example, wants to lay things out in specific places.
If any
A good friend
from a community shared by many on this list
often signs his e-mail WIRDI, whatever is right, do it.
I'm trying to get a handle on the /lib and /lib64 mess.
At first, I was put off by SuSE's use of /lib64, but it may be
the right thing. What do y'all think? What precedent
Yes, you could do that. The NSS could include a R/W or R/O ramdisk if you
wanted. ...
I encourage people to experiment with this support and to come to a
consensus as to how it can best be used. This was just a first step in
getting Linux for zSeries to use z/VM's proven shared-memory
I was wondering if anyone else had any experiences installing and
running Patrol under Linux/390, specifically how can it be installed so
Umm ... yes! But it has been a while. (early adopter here) ;-)
it can be shared among Linux images. It seems to want lots of
directories in
Per Jessen said:
Guys, you've got to realise that using Linux does in no way tie you in
with any particular vendor. ...
This is part of what we mean when we say
free as in speech -vs- Linux NOT being free as in beer.
There are costs, but the customers hands are UNtied.
And while we're at it, it would be helpful to have a userspace
program to do it and return a completion code (and maybe even do
the CP LINK as well).
I would disagree about the CP LINK part.
More significantly, this may be an ioctl() type of call.
Punching strings to /proc pseudo
Come on Rick. That isn't true with CMS either. Provided you save
Well ... I admit, you got me there:
CMS is locked in to 190, 19E, with a fondness for 191
and mild affinity for 19D and 192.
the NSS early enough it will be before the kernel finds out the
virtual machine size etc. If you
To make the bootstrap decide which ones are overrides
and which are additions is going to be ugly. I'm sure we
The bootstrap does not decide that.
could make it very complicated, but we decided to simply
append the parameters from the IPL statement to what is
Yes, a simple append is
[Wake up, Rick! Gabe did a cross post.]
Sorry, y'all, that I did not catch that and copy both lists.
-Original Message-
From: Rick Troth [mailto:rtroth;bmc.com]
Sent: Tuesday, October 22, 2002 5:04 PM
To: VM/ESA Discussions [EMAIL PROTECTED]
Subject: Re: Aduva?
Aduva is alive
Can we quit the OS-wars finger pointing, please?
Prior releases of MVS, VSE, VM, and so on did require a reboot
to change the clock. They also tended to use local time on the
HW clock. (We do now all use UTC on the HW clock, don't we??)
(Rhetorical question there, folks.) But that was
(cough) I'm sorry. My hearing is bad. You did what? (For a minute there
I thought you you modified PROFILE EXEC. Whew!) You meant to say that
you use the :vctc. tag in SYSTEM DTCPARMS to cause VM TCPIP to reestablish
the connections, right? Because we all know that you never modify
What Mark said. But also:
I recommend the reachable directory method.
(That's what SuSE calls it; I forget what RedHat calls it.)
In this method, you mount the CD manually one of three ways ...
as an FBA (if you can swing that! good luck!)
by way of loopback if you have
I have tried xterm display:192.168.1.10:0.0 but it says it can't open
display on 192.168.1.4
Try
xterm -display 182.168.1.10:0
When its done, you'll be in CP READ and you or anyone else will be able to
issue CP IPL system_name to start up your shared system. One caveat.
Every virtual machine that ipls your shared system must have the same disk
configuration as the system that was saved. That is, the disks must be at
NSS is Named Saved System.
You can take a snap-shot of a running (or runnable) system on VM
which CP (the hypervisor part of VM) will store into a spool file.
You can then IPL that system by name, rather than boot by device.
The syntax of the IPL command is (gross simplification)
I'm trying to bullet-proof my kernel maint.
Right now, I see the following dependency cascade:
linux-2.4.19 requires
glibc-2.2.5 requires
gcc-3.2 requires
binutils-2.12.90.0.15 requires
gettext-0.11.5
I stumbled
I've built Linux-2.4.19 with NSS support and the notimer patch, and made
it into a series of Debian kernel packages.
They're at http://www.sinenomine.net/downloads
This is GREAT!!!
On Fri, 8 Nov 2002, David Boyes wrote:
Much as I dislike Solaris, their diskless workstation filesystem layout
is a pretty good model for this. We should use that as a model for
ideas.
They also demonstrated the first shared /usr implementation.
They also do something I call folding (for
If you use the cmsfs stuff, that information can all be on the
191 disk and read by the startup scripts.
What about a CMSFS that can do directories and specials (device files)
akin to the UMSDOS hack?
On Fri, 8 Nov 2002, Kris Van Hees wrote:
Despite what Sun Microsystems did with linking /usr/bin and /usr/sbin
into the root filesystem as /bin and /sbin, a more sensible setup is
still to have the core utilities that are required to boot a system
(and to do basic maintenance) as part of the
If you create a CMS file called PROGRA~1 DIR I'll have to murder you.
;-)
Just so you know. Other than that, sure, sounds like a plan--I assume
you mean that you use some filesystem convention like a file which
always has some particular name, which contains a CMS filename to Unix
This folding, as far as I know, is just a couple of symlinks, from /bin to
/usr/bin and from /lib to /usr/lib. Doing the same thing on a typical Linux
Specifically, running 'ls -l' in root, you see
bin - usr/bin
lib - usr/lib
If memory serves, you do NOT
I have SSH installed on Linux, but have the opposite problem
in that I need an open ssh putty-like shell to connect to it
from OS/390 so that I can do secure logon from an automation product.
SSH was ported to OMVS ... I thought.
Can't seem to find it on the Tools and Toys page. :-(
Also--where are the current cmsfs kernel patches?
The ones at ftp.bmc.com don't build cleanly against 2.4.19.
The UTILITY builds fine. I just rebuilt it last week on
PC Linux, mainframe Linux, Solaris, HP (maybe even AIX).
The DRIVER I am working on.
(I thought I had it working for 2.4.7,
http://www.fsf.net/~adam/Mac-on-S390-desktop.png
Umm... Am I understanding that correctly?
MacOS on a pseudo-Mac hosted by Linux/390?
Put the disk and the ROM in a directory. Unpack the Basilisk sources.
From the top level, cd src/Unix; configure; make
Then tweak .basilisk_ii_prefs to
According to /var/run/gdm.pid, gdm was already running
But seems to have been murdered mysteriously.
GDM is the text mouse thingy.
Nifty if you run in text mode on a PC. Useless on the mainframe.
You can probably disable GDM with 'chkconfig'.
-- RMT
GDM is the graphical display manager for Gnome and it
used for graphical login purposes.
Ooopppsss... My bad.
I was thinking GPM. Sorry.
-- RMT
Attached patch fixes userland cmsfs; ...
Thanks!
I'll fit that in and check the kernelland effects.
-- RMT
I was looking for something else
and stumbled onto this post from the end of July
which I had missed at that time. (Too much good info on this list!)
To fix a collision between util-linux and s390-tools,
Karsten Hopp said:
The correct way would be to edit the util-linux.spec file
and rebuild
DEB (Legatus Tux) will remember this ...
In a previous lifetime, I learned the value of I18N.
The good doctor had our student employees coding up 'XMITMSG'
in REXX EXECs on the CMS systems. I hated it! Oh the pain,
oh the convoluted work, just to effect an echo. [sigh]
But I learned.
Depends on what YOUR virtual machine has set for line end.
The VM default here is #, but to be robust you really should
check with an 'hcp q term' before doing the 'hcp xautolog'.
The problem with # is that all UNIX shells use that to
introduce a comment and it need not be at start-of-line.
I'm trying to cobble-up a staged [re]build of the system,
based initially on the tool chain as indicated on DeveloperWorks.
Most of it works alright, but I keep bumping into a couple of errors.
Hard to tell if this is a chicken-and-egg scenario, but it doesn't
*look* like that. Most or all of
The glibc error seems familiar. I think I got it when I was trying to
compile glibc 2.2.5 with gcc 2.95.2/3.
Yeah ... good suggestion. This is 3.2.
I've brought the latest BINUTILS, GCC, GLIBC
as well as more than a dozen supporting packages. All very current.
The GLIBC error looks almost
Perhaps so, but if I were you, I'd visually inspect the
errlist.c file, just to be sure.
I did.
Swapping the strong and weak declaration *seems* to let it build.
But I then ran into other errors, so it will be some time before I
will know if that fixes it. Also, I really had hoped to do
BINGO! Mark ... you're a genius. Thank you.
Hmm. If the weak declaration is still in that file, then the patch was
_not_ applied. Here's what the section of code looks like on my system
after all the patches are put on:
Well ... I saw the code for myself, and it doesn't match yours.
So
So, has anyone considered this?
CKD volumes on the VM side make it tricky. (the usual situation)
If you have a wholly FBA based VM system (sadly, a rarity these days),
then backing up VM volumes via Linux is trivial. On FBA, data blocks
are always the same size and can be safely stored as
There is a CMSFS 1.1.6 on the BMC FTP site which incorporates
Richard Hirst's patch. Utility mode (user-land) seems fine.
Kernel-land (the driver) is still not reliable for other reasons.
ftp://ftp.bmc.com/pub/cmsfs/cmsfs-1.1.6.tar.gz
LeMarr, ...
The problem is not that you're on a 3270 session
but that on that 3270 session you're getting line-mode interaction
with the guest operating system.
Linux/390 now has (thanks to the kind folks at UTSGlobal)
a 3270 driver. Sadly, none of the full-screen text-mode applications
have
1 - 100 of 597 matches
Mail list logo
