Re: Debian Release Question
On Fri, Sep 10, 2004 at 01:05:22PM -0600, Hodge, Robert L wrote: When I look at packages at Debian web site, I see designations of woody4 or woody6 appended to the packages names. I know that woody is the codename for Debian release 3. What is the meaning of the numeric; i.e. 4 6 in this example? That construct is used to form version numbers for updates to the stable release. And how do I determine which woodyN version of a package I should install? You should install the version that apt-get install package gives you (the latest version available from stable). -- - mdz -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Adoption of UML Copy-On-Write
On Fri, Jun 18, 2004 at 01:50:48PM -0700, Brandon Darbro wrote: I have been looking into ways to add a writable layer to read only DASD for a while now, including ovlfs. Ovlfs simply doesn't do the job, unfortunately. I have problems with data getting lost between boots and it seems to continue to consume memory until the linux host eventually dies. So ovlfs is out. That brings me to User Mode Linux's layered block devices and it's Copy-On-Write feature. This appears to do exactly what I'm looking for, but is only available for the User Mode Linux virtual architecture. Any s390 kernel programmers here know if this could be implimented? Or are things just too different. What you're looking for are EVMS or LVM2 snapshots using device-mapper. -- - mdz -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Adoption of UML Copy-On-Write
On Fri, Jun 18, 2004 at 04:42:43PM -0700, Brandon Darbro wrote: Huh? EVMS or LVM2 has a method of adding a writable layer to read only dasd? EVMS supports writable snapshots, using copy-on-write from an EVMS volume. I haven't tried it with read-only DASD, but in theory it should be possible for it to be used this way. The act of creating the snapshot might require writing to the original volume in order to update the volume metadata. Ask on the EVMS mailing list. This is quite a lot closer to what you want than looking at UML; UML's COW feature is part of the ubd driver and is quite inapplicable to non-ubd devices, and what's more, UML doesn't work on s390. -- - mdz -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: SuSE vs Red Hat
On Thu, May 27, 2004 at 09:13:09AM -0700, Wolfe, Gordon W wrote: Debian has a reputation for being closer to the cutting edge of newest technology. I haven't worked with it, so I can't speak much to it. It does? Debian releases are less frequent than other vendors, so on the basis of individual software revisions it tends to be older. -- - mdz -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Some basic Debian-s390 questions
On Fri, Feb 06, 2004 at 01:17:20AM -0500, Alex deVries wrote: I have a few questions accumulated, although I'm sure there will be more. I suspect they're in the category of either: A - this is a really stupid question, or B - we've already answered this 1,000 times My questions so far are: 1. Problem installing When I went through the Debian installer, I got to a step labelled something like 'Install Kernel and Modules', which didn't. The installer segfaulted, and there was no kernel at all in /target/boot. The only way I could proceed was to manually locate the kernel .deb, transfer it and install that, and then rerun the installer, carefully skipping that step. Is this a known problem, or is it worth my time to reproduce? I've never seen this happen, but it's possible there is a bug. The installer is being completely rewritten for the next release, though, so I can't say whether this bug would be fixed, especially if it's not reproducible. 2. My apt sources Whenever I do anything with apt-get, I get a stream of complains that look like: W: Couldn't stat source package list http://security.debian.org stable/main Packages (/var/lib/apt/lists/security.debian.org_dists_stable_main_binary-s390_Packages) - stat (2 No such file or directory) I get one for stable/main, stable/contrib, stable/non-free, and then the whole listing again. What should be in my /etc/apt/sources.list file? You probably also get a message saying to run apt-get update, which is the solution to that problem. Those messages mean that you have certain sources specified in /etc/apt/sources.list, but haven't downloaded the lists of packages available from those sources (which is what apt-get update does). 3. fdasd didn't work fdasd didn't work out of the box because it was an old version. I have built a s390-tools_1.2.4-1_s390.deb package, how do I get it out there? I'm not sure I understand the question. If you're asking how to install a .deb, you should use dpkg --install. Did you build this package from source, or did you download it from testing or unstable? In the latter case, it is unlikely to install on woody, and needs to be recompiled. 4. Getting hipersockets working in the installer There's some mention of setting up a second ramdisk with the OCO drivers, but I never actually found the script to create such a disk. Given that there are now open source drivers for this, is there a replacement installer kernel, and where're the build instructions for the install kernel? I believe the instructions for the ramdisk are in the redbook. Information about replacing the installer kernel is here: http://www.debian.org/releases/stable/s390/ch-boot-floppy-techinfo.en.html#s-rescue-replace-kernel -- - mdz
Re: Some basic Debian-s390 questions
On Fri, Feb 06, 2004 at 10:52:42AM -0500, Alex deVries wrote: Matt Zimmerman wrote: You probably also get a message saying to run apt-get update, which is the solution to that problem. Those messages mean that you have certain sources specified in /etc/apt/sources.list, but haven't downloaded the lists of packages available from those sources (which is what apt-get update does). So I ran 'apt-get update', and that too gave me the same stream of errors. What's next? It almost certainly gave a different stream of errors, about failing to download the package lists, rather than not finding them. I'm not sure I understand the question. If you're asking how to install a .deb, you should use dpkg --install. Did you build this package from source, or did you download it from testing or unstable? In the latter case, it is unlikely to install on woody, and needs to be recompiled. Let me rephrase: I couldn't find an s390-tools_1.2.4-1_s390.deb package out there anywhere, so I created and built my own. I'm sure there are other people who have had this problem too. Would anyone like this package? Probably; I would send a message to [EMAIL PROTECTED] -- - mdz
Re: Some basic Debian-s390 questions
On Fri, Feb 06, 2004 at 10:05:22AM -0800, Matt Zimmerman wrote: Probably; I would send a message to [EMAIL PROTECTED] No sooner did I speak than 1.2.4 was uploaded to unstable. -- - mdz
Re: compile failures for the 2.4.21 and 2.4.23 kernels
On Thu, Feb 05, 2004 at 11:10:37AM -0500, Alex deVries wrote: Debian does indeed have a lot of goodness. If I were interested in asking some relatively basic Debian-s390 questions (like: how do I update my s390-tools so that fdasd actually works?), would this be the right place? Or should I ask them on the debian-s390 list? The debian-s390 list would be the best place for s390-specific Debian questions. For general Debian questions: http://www.debian.org/support -- - mdz
Re: Kernel 2.4.24 available for SLES8?
On Wed, Jan 14, 2004 at 03:03:25PM -0600, Little, Chris wrote: there are several of vulnerabilities that are x86 specific. the one that immediately comes to mind is the one that plagued the debian servers. If you mean CAN-2003-0961, that vulnerability is not i386-specific, though it only affects a subset of Linux architectures. -- - mdz
Re: Setting up an APT Server
On Thu, Jan 15, 2004 at 12:36:31AM -0500, Mark Post wrote: Hopefully one of the more Debian-literate people on the list can point me in the right direction... Every SHARE, we have a double installfest on Friday morning. We take 30 or so Windows systems, and install Debian and then Red Hat on them. Everyone seems to really like the hands-on aspect, but one problem we have is that part of the installation process asks if it should check for updates. It would be nice if we had a local-to-the-lab APT server so that we could let them see what that looks like, without killing the T1 connection. So, can anyone point me to a step-by-step procedure on how to set that up? I'm not terribly Debian literate myself, so it would have to be pretty detailed. If your concern is really bandwidth, it would be much simpler to just set up squid and have everyone proxy through it; that way you only download each package once, which works out quite well for a large number of near-simultaneous network installations. Then, you wouldn't have to worry about setting up a local package repository at all. But, to answer your question, the tool that you want is apt-ftparchive(1), included in the apt-utils package. It has a detailed man page, but for the simplest case, a flat directory full of random .debs for a single architecture, the usual command line would be: apt-ftparchive packages . | gzip -9 Packages.gz and the corresponding line in /etc/apt/sources.list would look something like this: deb http://some.host/path/to/dir ./ -- - mdz
Re: Anyone Nagios? (GPL discussion)
On Tue, Jan 06, 2004 at 10:33:41AM -0800, Fargusson.Alan wrote: Nope, never heard of them. OK, just kidding. What I meant is that products derived from these have not lasted. BTW: I am not sure that X counts, since Xfree86 is not based on X source code. XFree86 is licensed under a BSD-like open source license (just like X), not the GPL. -- - mdz
Re: More on Fedora...
On Tue, Jan 06, 2004 at 04:16:32PM -0500, Alex deVries wrote: The one thing I find a little disappointing of late is that there's no distribution that is: a) RPM-based b) for s390 c) kept up to date d) free as in beer I can get three out of four of those (Debian (bcd) , SLES8 (abc) and RHAS3 (abc)), but not all four. Fedora would hit all four if it were built for s390. Has anyone looked at this yet? I've gotten the appropriate cross compilers working, and was having some problems with building glibc. Why is RPM a requirement? This isn't exactly i386 after all, so it's not as if there are bunches of third-party binary packages floating around that you can take advantage of. -- - mdz
Re: How to make the 'mail' command talk to MS Exchange Server
On Tue, Dec 30, 2003 at 03:50:13PM -0600, McKown, John wrote: There is some discussion about not wanting SendMail due to it's many exploits. Also, the thought was why implement an MTA when we're not receiving any mail, just sending it. Sending is transport too. If you want the barest of MTAs to connect to a relay and nothing else, try ssmtp. That will turn any /usr/sbin/sendmail-using program into the sort of thing that you want, including /bin/mail. Make sure that's what you want, though; sometimes you want a real MTA even if you don't think that you do. For example, ssmtp doesn't even have a local queue; if it can't talk to the relay, it will fail immediately. -- - mdz
Re: Patch to glibc
On Tue, Dec 16, 2003 at 01:15:05PM -0500, Mark Post wrote: Mmmm, no, I don't think so. I really want the Linux/390 folks in Boeblingen to figure out what would be the correct fix for this. After I sent the patch, I re-read Ulrich's post much more closely, and re-discovered his thought that the correct way might be to introduce another makefile variable to handle this value for just the .oS files. That sounds fine. Do the folks in question read this list? I don't see much development traffic in these parts. I would try the email address listed in the code in that directory: /* Copyright (C) 2000, 2001 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Martin Schwidefsky ([EMAIL PROTECTED]). -- - mdz
Re: Patch to glibc
On Mon, Dec 15, 2003 at 08:12:22PM -0500, Mark Post wrote: I don't know, and cannot test, if a similar fix needs to be applied to glibc-2.3.2/sysdeps/s390/s390-64/Makefile or not. Someone who is running on s390x might be able to check it out and let the rest of us know. If this is _not_ the correct fix, I would appreciate knowing what is, so that I can modify my build scripts appropriately. If you haven't already, you might forward this to [EMAIL PROTECTED] or such to remind the glibc maintainers about it. -- - mdz
Re: Getting gcc Ada to build
On Sun, Nov 23, 2003 at 05:43:50PM -0500, Post, Mark K wrote: Well, that raises the question of how Debian and Red Hat got an Ada compiler to start with. Or for that matter, Slackware, or any of the other distributions out there. Was there some free Intel-based Ada compiler (non-gcc) available that everyone used? It was presumably bootstrapped at some point the same way that gcc was, using a cross-compiler. Once binaries exist, they're self-sustaining. -- - mdz
Re: Linux ready for the desktop: IBM
On Sun, Nov 16, 2003 at 07:30:39PM +0800, John Summerfield wrote: On Fri, 14 Nov 2003, Matt Zimmerman wrote: On Thu, Nov 13, 2003 at 08:40:23AM +0800, John Summerfield wrote: IMV it would be a Good Thing for IBM to ensure all its Windows desktop software runs under WINE. I'm sure the WINE project would welcome the input, and I'm sure that IBM has the expertise to resolve any problems, whether by changing WINE or by changing its own software. I'm sure there's a Microsoft lawyer somewhere salivating over that possibility. Why? Given that Microsoft's partners have been restricted from even _distributing_ competing software under other circumstances, I would assume that they also considered the possibility of developing software to directly challenge their market position. -- - mdz
Re: bash and zsh question
On Wed, Nov 12, 2003 at 09:06:55AM -0500, Froberg, David C wrote: Curious if anyone has tried shifting from using bash to zsh. If someone did switch or tried it and gave up, I would be curious to know what you encountered. I didn't switch from bash, but I've been using zsh for some time now and never looked back. -- - mdz
Re: Linux ready for the desktop: IBM
On Thu, Nov 13, 2003 at 08:40:23AM +0800, John Summerfield wrote: IMV it would be a Good Thing for IBM to ensure all its Windows desktop software runs under WINE. I'm sure the WINE project would welcome the input, and I'm sure that IBM has the expertise to resolve any problems, whether by changing WINE or by changing its own software. I'm sure there's a Microsoft lawyer somewhere salivating over that possibility. -- - mdz
Re: OT: New Mac G5
On Thu, Oct 30, 2003 at 07:41:24PM -0600, Jay Maynard wrote: On Thu, Oct 30, 2003 at 06:47:06PM -0500, Matt Zimmerman wrote: Hercules at least builds on all Debian architectures, including a couple which fit that profile (sparc64 and ia64), and has for some time. I've only actually run it on sparc64 in 32-bit mode, and didn't do much with it. True. However, the ia64 doesn't help here, since it doesn't execute 32-bit code natively. Are you sure about that? The SPARC64 might be an interesting case, though. What I'm looking for is Hercules built as a 32-bit and a 64-bit app, for the same architecture, running 64-bit Linux. I think the 64-bit host will show a nice speed improvement, but that's just a gut feeling at the moment. You could perform that test on either architecture. -- - mdz
Re: OT: New Mac G5
On Mon, Nov 03, 2003 at 09:06:59AM -0800, Fargusson.Alan wrote: I will concur with Jay here. The IA64 is a new instruction set that is not compatible with the IA32. This is a much different situation than exists with SPARC and zSeries, which are the same instruction set with additional instructions for 64 bit. You are correct in that ia64 is a new instruction set which is not a superset of ia32. However, Itanium 2 implements both ia64 and ia32 natively. ia32 performance is less than Intel's top of the line ia32-only processors, but I don't know how it compares to those with similar clock speed (or some other metric). -- - mdz
Re: OT: New Mac G5
On Thu, Oct 30, 2003 at 02:20:51PM -0500, Guillaume Morin wrote: Dans un message du 30 Oct ` 13:55, Arty Ecock icrivait : I'm considering replacing several existing development workstations with a single G5 Mac. The G5 would run Linux natively with the following applications: You have to know that the port is not yet complete and that for now it is really suboptimal. You mean support for the G5 specifically, and not the powerpc port in general, right? -- - mdz
Re: OT: New Mac G5
On Thu, Oct 30, 2003 at 03:14:23PM -0600, Jay Maynard wrote: The only difference, I expect, will be that Hercules on the G5 will run 64-bit mode faster, since it won't have to execute multiple instructions to do single 64-bit operations. I don't know how much difference this will make, however, and benchmark results would be VERY welcome. (At present, there is no architecture that I know of to which Hercules has been ported that executes both 32- and 64-bit code natively. Well, except 390/z/Series, but I don't know of anyone who's run a comparison there.) Hercules at least builds on all Debian architectures, including a couple which fit that profile (sparc64 and ia64), and has for some time. I've only actually run it on sparc64 in 32-bit mode, and didn't do much with it. -- - mdz
Re: GateD for Linux/390 SuSE 7?
On Tue, Oct 28, 2003 at 10:29:18AM -0500, David Boyes wrote: IBM wants us to see if we can get the Linux to send RIP2 packets, which I believe only can be done with gateD. Is there an rpm available for SuSE 7 Linux? Use Zebra instead of gated. It supports RIP2. Use GNU Quagga instead of Zebra, as Zebra is dead and Quagga rose from its ashes. -- - mdz
Re: OT: Intel gets virtualization clue?
On Fri, Oct 10, 2003 at 12:11:32PM -0400, Paul Hanrahan wrote: Jan, [...] Have you tried running Hercules ? Yes, I think that Jan may have some experience in that area. ;-) -- - mdz
Re: qeth module
On Thu, Oct 16, 2003 at 10:52:13AM -0500, Dennis Wicks wrote: It appears that I need to install the qeth module on my Debian/Woody 2.4.19 system. How do I do this? http://www.debian.org/releases/stable/s390/ch-hardware-req.en.html#s-network-cards -- - mdz
Re: Rsh logon error when routing to another server
On Fri, Sep 26, 2003 at 12:27:10PM -0600, Nakagawa, Robert wrote: Using Debian z800 version 3.01r needing help to resolve a problem with using the command rsh to another server. The client asks for a password when the remote server does not require a password. I have installed ssh on the client and this seems to be the problem. How do I get the standard rsh binary back instead of the rsh-ssh binary. Next I have remove the ssh package and this did not help, still had the same problem. Probabl you do not have rsh (in the rsh-client package) installed. Because ssh is command-line compatible with rsh, and is often used with programs which expect this interface, ssh provides the alternative /usr/bin/rsh. If you install rsh-client, that client will provide a higher-priority alternative for /usr/bin/rsh. You can examine the current alternatives with update-alternatives --config rsh. -- - mdz
Re: Any new SuSE vs RedHat arguments?
On Fri, Sep 19, 2003 at 12:40:16PM -0500, Chris Cox wrote: Actually, not unfair at all. Though as a Debian person you are quite used to doing many things on your own outside of support, corporate customers are not likely to install packages outside of the vendor's support. Actually it is quite the opposite. I tend to find everything that I need already provided and supported by Debian and need to install very little additional software. -- - mdz
Re: Any new SuSE vs RedHat arguments?
On Fri, Sep 19, 2003 at 02:56:59PM -0500, Chris Cox wrote: Matt Zimmerman wrote: Actually it is quite the opposite. I tend to find everything that I need already provided and supported by Debian and need to install very little additional software. Good for you. My experience is that Debian is a bit out of date. I'm guessing your running Debian unstable (?). That's your choice. Just somewhat of a moving target. Debian releases infrequently. Depending on your business and system practices, this can be an advantage or a disadvantage. I run stable releases on servers, and unstable on some desktops. In both cases, I generally get what I want from it. In the exceptional circumstances where I require a specific feature or bug fix, I upgrade individual packages on my stable systems. I'd be curious to hear what you found to be missing from Debian 3.0. Most of the complaints I hear tend to boil down to version numbers are smaller, or misunderstandings about backported security fixes (which actually boils down to the same issue of version numbers). -- - mdz
Re: PDF version of Installing Debian GNU/Linux 3.0 For S/390?
On Tue, Sep 09, 2003 at 12:34:54PM -0500, McKown, John wrote: I really want to print this manual out and read it. Is there a PDF version available? I've found that printing Web pages just is not very high quality. If not, is there a step by step guide that I can print out? I've got a Debian/390 installation at home under Hercules/390. But I don't really know where to go from here. I want install packages et al. I do have a fair understanding of how to do this sort of thing using RedHat on Linux/Intel. But I'm not familiar with Debian at all. There is a link to the PDF version on the same page as the link to the HTML version: http://www.debian.org/releases/stable/installmanual However, if you're looking for information on how to do the usual sorts of things once the system is up and running, the installation manual probably isn't the right document. Try the APT HOWTO: http://www.debian.org/doc/manuals/apt-howto/ There is a PDF version of that one as well, which is linked from this page: http://www.debian.org/doc/user-manuals (which is an index of much of the Debian documentation available) There is also this short document that I wrote: http://people.debian.org/~mdz/debian-for-unix-users.html/ which is a crash course in the Debian packaging system aimed at users who are familiar with Unix-like operating systems and, ideally, packaging systems in general as well. There is not currently a PDF available, but if I ever get around to working on it again, I can easily generate one from the SGML source. -- - mdz
Re: gcc 3.3 internal compiler error
On Mon, Sep 08, 2003 at 12:19:36PM -0400, Post, Mark K wrote: I'm trying to compile Mozilla with gcc 3.3. I'm getting the following ICE, on the same module, every time. Has anybody else seen this, or know of a patch that will correct it? I can say that the gcc 3.3.1 (+patches) in Debian unstable builds the Mozilla 1.4-3 Debian package without problems. You may want to look at the Debian patches to gcc and the Debian patches to mozilla. The ICE is definitely a compiler problem, but there are also s390-specific patches to mozilla which you will probably need. http://buildd.debian.org/fetch.php?pkg=mozillaver=2%3A1.4-4arch=s390stamp=1063019704file=logas=raw [...] nsAFMObject.cpp g++ -o nsAFMObject.o -c -DOSTYPE=\Linux2.4\ -DOSARCH=\Linux\ -I../.. -I./.. -I../../../dist/include/xpcom -I../../../dist/include/string -I../../../dist/include/widget -I../../../dist/include/pref -I../../../dist/include/caps -I../../../dist/include/locale -I../../../dist/include/uconv -I../../../dist/include/view -I../../../dist/include/necko -I../../../dist/include/imglib2 -I../../../dist/include/gfx -I../../../dist/include -I/build/buildd/mozilla-1.4/build-tree/mozilla/dist/include/nspr -I/usr/include/freetype2-fPIC -fno-rtti -fno-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-long-long -O2 -DDEBIAN -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -DDEPENDENT_LIBS=\libgkgfx.so\, \libxpcom.so\, \libplds4.so\, \libplc4.so\, \libnspr4.so\, \libpthread.so\, \libdl.so\, \libc.so\, \libpthread.so\, -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/pango-1.0 -I/usr/include/freetype2 -I/usr/X11R6/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -DMOZILLA_CLIENT -include ../../../mozilla-config.h -Wp,-MD,.deps/nsAFMObject.pp nsAFMObject.cpp n [...] -- - mdz
Re: gcc 3.3 internal compiler error
On Mon, Sep 08, 2003 at 02:55:36PM -0400, Post, Mark K wrote: Thanks. I do have the patches on for Mozilla. The build ends very quickly if you don't. Can you give me a direct pointer to the gcc 3.3.1 patches? http://any debian mirror/debian/pool/main/g/gcc-3.3/gcc-3.3_3.3.2ds1-0pre2.diff.gz (http.us.debian.org if you don't know of anything closer) -- - mdz
Re: How do I build vmlinuz manually?
On Fri, Sep 05, 2003 at 08:07:37AM -0500, Adam Thornton wrote: On Thu, 2003-09-04 at 23:24, Matt Zimmerman wrote: make zimage (note the case, different from i386) should do it. Maybe it should, but in the top-level directory: debian:/devel/kernel/linux-2.4.21-plus# make zimage make: *** No rule to make target `zimage'. Stop. Which is why I'm wondering what the incantation to turn vmlinux into vmlinuz by hand is. Or maybe I'm asking the wrong question, and what I really should be asking is why can't I IPL the vmlinux that 'make boot' produces, but I can IPL the vmlinuz that 'make-kpkg buildpackage' produces? As someone else mentioned, make image is correct. I looked to see what make-kpkg was using, but misread it: ifeq ($(strip $(architecture)),s390) kimage := zimage loaderdep=zipl loader=zipl loaderdoc= target = image NEED_DIRECT_GZIP_IMAGE=NO kimagesrc = $(strip arch/$(KERNEL_ARCH)/boot/$(target)) kimagedest = $(INT_IMAGE_DESTDIR)/vmlinuz-$(version) DEBCONFIG= $(CONFDIR)/config.$(KPKG_SUBARCH) endif -- - mdz
Re: Is there a way to get scp to read STDIN?
On Thu, Sep 04, 2003 at 12:19:05PM -0500, McKown, John wrote: I want to do something like: program | scp - [EMAIL PROTECTED]:/output.file program | ssh [EMAIL PROTECTED] 'cat output.file' -- - mdz
Re: How do I build vmlinuz manually?
On Thu, Sep 04, 2003 at 10:39:16PM -0500, Adam Thornton wrote: So, I'm building my own 2.4.21 kernel, with the IBM patches and some of Leland Lucius's stuff. make or make boot builds vmlinux. There's no obvious compressed target in the arch/s390/boot directory. And when I try to boot that kernel, I get a program interrupt loop. But when I use the Debian tools (make-kpkg), I end up with a vmlinuz that boots fine. So my question is, what operations do I need to perform on vmlinux to turn it into vmlinuz myself? make zimage (note the case, different from i386) should do it. -- - mdz
Re: Merge SuSE updates and original CDs???
On Fri, Aug 29, 2003 at 09:19:28PM +0300, Tzafrir Cohen wrote: Anyway, debain has no formal mirrors of the security apt source (unlike , say, Mandrake) http://www.debian.org/security/faq explains why. Note that, of course, no one is prohibited from mirroring the archive if it is convenient for them; they simply accept the risk of updates not being available promptly if there is a problem with the mirroring. -- - mdz
Re: Linux and the P/390
On Thu, Aug 28, 2003 at 07:05:16PM +0100, Alex J Burke wrote: Lastly does anybody know how to configure two (in future 3) LCS interfaces, one being Ethernet the other Token Ring? I set the lcs.chandev, set the lcs, set the the s390 architecture file, added the extra IP to hosts and added tr0 to the interfaces file. I then did a modules-update (all modifications were done to the /etc/modutils, /etc and /etc/network repectivelythe correct places). The tr0 still refuses to come upeth0 however always comes up. An extra set of 3088s were defined in the P/390 configurator for the TR interface, the 802.2 protocol was added to the TR interface under OS/2.it was the only protocol on the two mainframe connections. Any thoughts? Try running ifup tr0 manually. If the interface comes up OK, you probably forgot to use the auto keyword to configure it to come up automatically at boot. If not, look for error messages to find out why. -- - mdz
Re: SuSE and shortcut keys
On Wed, Aug 06, 2003 at 10:11:00AM +0200, Ceruti, Gerard G wrote: Another question , is there an idiots guide to all the doco available via man ?, If your apropos(1) program supports regular expressions (check man apropos), then you can get a list with apropos .. Of course, this list will be enormous on a system with much software installed, and it isn't a good way to find documentation. Use apropos(1) to search for keywords or regular expressions instead. -- - mdz
Re: zSeries performance heavily dependent on working set size
(would it be possible for you to configure your mail software to add attribution lines when quoting? it makes it much easier to follow who said what) On Mon, Aug 11, 2003 at 05:22:11PM -0700, Jim Sibley wrote: I think that the word you're searching for is pathological. Its only pathological in the sense that is forces the condition to emphasize its effect. We did not start with this case. It's still pathological, even if it was done intentionally (perhaps especially so). What we did notice is that a data base random write was much slower than a data base random read for a large data base. Examination showed that no I/O was actually taking place because of the Linux caching of buffers and retention of dirty buffers. (and as I said before, no paging). We first eliminated the I/O entirely by reading/writing to /dev/shm and saw the same effect - the random reads were much slower than the sequential reads to the data base. First you were comparing random write to random read, then random read to sequential read. Is this correct? That would seem like (potentially) a different problem. Throwing more hardware at it won't solve this poor performance problem. Adding CP's or memory may aggravate it! This is not generally true, though it may be in a consolidated scenario. For example, if your working set is 1GB, and you can efficiently segment that working set and forward requests to N physical servers rather than a single one, each of those servers will have a smaller working set and be more efficient in itself. If these are instead virtual images, the effect is less predictable (at least for me). The only solution that I see is reducing locality of reference (working set size). Reducing working set size increases locality of reference (which is what you want). -- - mdz
Re: A Very Serious Virus Alert
On Tue, Aug 12, 2003 at 03:05:16PM -0400, Loren Charnley, Jr. wrote: I think that I have received this message enough, 3 times, in less than an hour. Please stop! Additionally, even one message would have nothing to do with Linux/390, Hercules, or whatever other mailing lists were spammed. -- - mdz
Re: zSeries performance heavily dependent on working set size
On Mon, Aug 11, 2003 at 10:06:03PM +0200, Rod Furey wrote: Erm... far be it from me to argue with the master on the performance of the actual physical hardware, but isn't the point here that they've generated a worst-possible-case scenario (is this called a degenerate case these days?) I think that the word you're searching for is pathological. -- - mdz
Re: assembler listing for Gnu compiler
On Wed, Aug 06, 2003 at 07:20:53AM +0800, John Summerfield wrote: On Tue, 5 Aug 2003, Jim Sibley wrote: man cc does not seem to work on SLES8. How do you turn on the assembler listing for the compiler (cc source.c ...)? Doesn't work on Debian either. However, man gcc does. It doesn't? With what version of the 'gcc' package? dijkstra:[~] man -w cc /usr/share/man/man1/gcc-2.95.1.gz dijkstra:[~] dpkg -l gcc Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name VersionDescription +++-==-==- ii gcc2.95.4-14 The GNU C compiler. -- - mdz
Re: Installing Debian 3.0r1 (woody) on Hercules/390
On Wed, Jul 30, 2003 at 08:46:44AM -0500, McKown, John wrote: I'm obviously doing something really stupid. I'm running RH 9.0 on my Intel box at home. I run Hercules/390 2.17.1 to create an S/390 environment. I ipl'ed the Debian 3.0r1 (woody) install on this. I get IP connectivity from the installation program to the RH 9.0 host. I can format the 3390-3 emulated DASD just fine. But when I try to install the kernel, I'm dead in the water. I tried using the ftp method. Apparently I have something set up wrong on RH 9.0 (vsftpd) because anonymous cannot see much of anything. I can't figure out this problem, I spent about 3 hours last night trying to configure vsftpd correctly and got nowhere. I noticed that the ftp method used wget. I could invoke a shell on the Debian install and found that if I used the form: wget ftp://userid:[EMAIL PROTECTED]//... that I could get to the files. However when I try to use that form in the installation dialog, the installer takes the userid as the hostname and dies. Hmm, I have not encountered this before, but I have never had cause to try to install Debian from a password-protected source. Since all of Debian is freely redistributable, there is normally no need to protect it from anyone. :-) You would probably have an easier time setting up Apache and using HTTP than setting up vsftpd. So, I tried using the NFS method. I set up the exports correctly. I know they work because I can use another Linux system on my LAN to do an NSF mount of the Debian installation directory on my RH 9.0 system. But all I get when I try on Debian is incorrect parameter or some such thing. Invalid argument. This is a known problem; NFSv3 does not work correctly with the installation kernel. http://www2.marist.edu/htbin/wlvtype?LINUX-VM.30073 You can work around the issue by using NFSv2 (mount -o nfsvers=2). Again, this was trying using the invoke a shell function on the Debian install, using the mount command. I plan to create a 3390-9 volume next. I will then IPL the installation tape, do an invoke a shell and use the wget command that I got working to copy the installation files from the RH 9.0 to the filesystem on the 3390-9 volume (they should all fit). That should work as well. -- - mdz
Re: Installing Debian 3.0r1 (woody) on Hercules/390
On Wed, Jul 30, 2003 at 08:57:14AM -0500, McKown, John wrote: See, I said it was stupid. Thanks, I'll change the home directory of ftp to / and try again tomorrow night (tonight is too busy). This system is not on the Internet, just my home LAN. The PC on the Internet has a paranoid firewall configuration, so I don't worry about that. You can accomplish the same thing without opening up your entire system to anonymous FTP by moving the files into ftp's home directory (or the web server root), or bind-mounting the directory there. -- - mdz
Re: Installing Debian 3.0r1 (woody) on Hercules/390
On Thu, Jul 31, 2003 at 02:27:29AM +0800, John Summerfield wrote: On Wed, 30 Jul 2003, Matt Zimmerman wrote: Hmm, I have not encountered this before, but I have never had cause to try to install Debian from a password-protected source. Since all of Debian is freely redistributable, there is normally no need to protect it from anyone. :-) RHL 9 is also freely redistributable: what you're controlling is access to the server, not to the data. No, you're controlling access to the data. Anyone can still connect to the server when it is password-protected; you're only controlling access to the data on it. And in this case, the server wasn't publicly accessible anyway, and the password protection only existed as a workaround for problems in setting up anonymous FTP. -- - mdz
Re: Stripping trailing blanks?
On Tue, Jul 29, 2003 at 06:30:32AM +0800, John Summerfield wrote: man perlfunc perldoc -f chomp -- - mdz
Re: Problems with sysstat package on Debian
On Fri, Jul 18, 2003 at 08:37:55AM -0600, Hodge, Robert L wrote: I'm having problems with the sysstat package on the Debian 3.01 distribution. sar -b is returning all zeroes for disk activity and iostat is not reporting anything for disk. It looks like iostat is not finding any disks. I've searched the archives of this list and other lists with no positive results. I've done numerous web searches with no results. I've upgraded the sysstat package to 4.1.2 from 4.0.4 with no resolution to the problem. I've upgraded the Linux kernel to 2.4.21 from 2.4.17 with no resolution to the problem. The Debian guest is running under z/VM 4.3 on a z/800 processor. Can anyone tell how to fix the sysstat package on Debian 3.01 for S/390? It looks like this is a kernel issue, and not systat. If you look in /proc/stat, you'll see that the disk_io line is empty, whereas on (for example) an Intel system with SCSI disks, it contains info like this: disk_io: (3,0):(454481,215862,18852900,238619,11184288) (8,0):(742898,347204,9634940,395694,6420320) I only tested with the 2.4.17 kernel in Debian 3.0; this may be fixed in newer kernels such as the 2.4.19 kernel in proposed-updates and unstable. -- - mdz
Re: Suse Version
On Tue, Jul 15, 2003 at 08:28:31PM -0500, Tom Duerbusch wrote: John Summerfield wrote: Please, take care with your language. public domain means that I can take it, change it, copyright it, not reveal my changes etc. Yes, you are right. I keep getting the variety of terms mixed up. Public Domain isn't the same as GNU (is it?). Correct. Those terms aren't even on the same plane. GNU is a project formed to develop software (specifically an operating system), while public domain is a legal term indicating the absence of copyright. In order to meet the goals of the GNU project, a license was developed, called the GNU General Public License (often abbreviated GPL), and most of their software is released under this license (copyrighted, and not in the public domain). You can learn more about it than you ever cared to know here: http://www.gnu.org/licenses/gpl-faq.html And it seems like there is a third type (or perhaps just another term that is the same as a previous term). If you mean licenses, there are too many to enumerate, but the terms which cause the most confusion in this area are free software and open source, both of which are described in excruciating detail in many documents which can be easily located on the web. -- - mdz
Re: which version of reiserfs?
On Tue, Jul 01, 2003 at 11:11:13AM -0400, Hall, Ken (IDS ECCS) wrote: SLES8/EXT3 (formatted with -T largefile4) -rw-r--r--1 root root 5100758784 May 27 13:01 bigfile The caveat is that with EXT2 or EXT3, you have to specify -T largefile4 or the filesystem is created with 1k/inode. largefile4 creates 4MB/inode, which wastes considerable space on small files, but allows files much larger than 2 gb. The 'largefile4' option is a only tuning parameter which can be used to gain space efficiency on a filesystem which whose average file size will be large. It is NOT necessary in order to allow the creation of files larger than 2GB, and will cause many headaches if used inappropriately, since with a normal file size distribution, the filesystem will run out of inodes very quickly. ext2 and ext3 support large files out of the box given appropriate userland support (which will be found in any modern Linux distribution). -- - mdz
Re: Cron sending e-mail
On Wed, Jun 25, 2003 at 02:55:56AM +0800, John Summerfield wrote: On Tue, 24 Jun 2003, Fargusson.Alan wrote: Try something like this: 10 * * * * root /usr/sbin/ntpdate -u condor /tmp/Log 21 No. 1. Logs in linux belong in /var/log 2. Each run clobbers the previous report. 3. Nobody will look at it. 4. Allows users to clobber arbitrary files as root via a symlink attack -- - mdz
Re: zSeries and Laptop
On Sat, Jun 21, 2003 at 02:53:00AM +0800, John Summerfield wrote: You don't need pm to do backups. You _might_ need it to resize your partition, but as far as possible use Linux tools. I resized a Windows 98 partition using FIPS which came with Red Hat Linux. I backed uo the whole drive first, using dd to copy it to a file across the LAN, gzip to compress the file and cdrecord (maybe mkiosfs too) to create a CD. GNU parted works nicely for this nowadays. It's included in Debian, and on Knoppix CDs. btw There's an open source program called User Mode Linux, hosted at sf.net. Like VM it lets you run several Linux instances simultaneously on one box. AFAIK it's only available for IA32, though someone was porting it to PPC. I used it to run Red Hat Linux 7.0 on Red Hat Linux 6.x. There is also a semi-functional port of UML to Win32, interestingly enough. -- - mdz
Re: Suse YOU updates
On Wed, Jun 18, 2003 at 10:46:12AM -0500, Tom Duerbusch wrote: Now I think I know what apt-get does. It seems to be a Debian version of RPM (for Redhat and Suse), which, I think, YaST is a nice shell around RPM (for Suse users). So apt-get and RPM are command line versions with YaST being an interactive frontend. apt-get is a bit higher-level than RPM; the RPM equivalent on Debian is dpkg. apt-get is a command-line front-end to dpkg. aptitude and dselect are interactive text-mode front-ends, probably closer to YaST (though I've never used it). -- - mdz
Re: invalidate: busy buffer While Booting
On Sat, Jun 14, 2003 at 06:19:05AM +0800, John Summerfield wrote: On Fri, 13 Jun 2003, James Melin wrote: Is there a way to convert to ext3 or reiserfs 'on the fly'? Or do you have to backup, reformat, restore? man tune2fs ...for safe and painless ext2-ext3 conversion, or for more adventurous experimental conversions, there is this interesting tool: http://members.optusnet.com.au/clausen/ideas/convertfs.txt -- - mdz
Re: any debian users
On Fri, Jun 13, 2003 at 08:53:51PM +0800, John Summerfield wrote: On Fri, 13 Jun 2003, Noll, Ralph wrote: zvmlinx7:/tsm# rpm -ivh TIVsm-BA.s390.rpm error: failed dependencies: /bin/sh is needed by TIVsm-BA-5.1.5-0 ld.so.1 is needed by TIVsm-BA-5.1.5-0 libcrypt.so.1 is needed by TIVsm-BA-5.1.5-0 libcrypt.so.1(GLIBC_2.0) is needed by TIVsm-BA-5.1.5-0 libc.so.6 is needed by TIVsm-BA-5.1.5-0 libc.so.6(GLIBC_2.0) is needed by TIVsm-BA-5.1.5-0 libc.so.6(GLIBC_2.1) is needed by TIVsm-BA-5.1.5-0 libc.so.6(GLIBC_2.1.3) is needed by TIVsm-BA-5.1.5-0 libc.so.6(GLIBC_2.2) is needed by TIVsm-BA-5.1.5-0 libdl.so.2 is needed by TIVsm-BA-5.1.5-0 libdl.so.2(GLIBC_2.0) is needed by TIVsm-BA-5.1.5-0 libdl.so.2(GLIBC_2.1) is needed by TIVsm-BA-5.1.5-0 libm.so.6 is needed by TIVsm-BA-5.1.5-0 libm.so.6(GLIBC_2.0) is needed by TIVsm-BA-5.1.5-0 libpthread.so.0 is needed by TIVsm-BA-5.1.5-0 libpthread.so.0(GLIBC_2.0) is needed by TIVsm-BA-5.1.5-0 libpthread.so.0(GLIBC_2.1) is needed by TIVsm-BA-5.1.5-0 libpthread.so.0(GLIBC_2.2) is needed by TIVsm-BA-5.1.5-0 libstdc++-libc6.2-2.so.3 is needed by TIVsm-BA-5.1.5-0 getting the above errors while trying to install TSM on a debian linux/390 31 bit running under z/VM any ideas or point me in a direction..hopefulle the right direction While you're waiting for mdz to comment, read up on Alien and try converting the rpm to a deb. rpm doesn't have any idea of what you have on your Debian system. Yes, that will probably work (though watch out for scripts which run at installation time). All of these libraries should be available on the system by default except perhaps for that older libstdc++, which is available in the libstdc++2.10-glibc2.2 package. If in doubt, you can use the form here: http://www.debian.org/distrib/packages to find which package contains a particular file. -- - mdz
Re: any debian users
On Fri, Jun 13, 2003 at 09:26:54PM +0800, John Summerfield wrote: On Fri, 13 Jun 2003, Sergey Korzhevsky wrote: or use rpm --nodeps if you have all needed packages. That's only for the desperate. --nodeps is almost never a good idea. Likewise for installing RPMs and Debian packages directly on the same system. :-) -- - mdz
Re: hz_timer
On Thu, Jun 12, 2003 at 10:24:08PM +1200, Vic Cross wrote: On Thu, 12 Jun 2003, John Summerfield wrote: On Wed, 11 Jun 2003, Rod Furey wrote: A more interesting question is: what do the User Mode Linux patches do? Run real, unmodified* Linux programs in a virtual computer. UML uses standard user-space API to access virtual devices and memory. Exactly, but I suspect that Rod was actually asking what UML does about the kernel timer... ;-) Let's just say that time inside UML is...different. :-) -- - mdz
Re: Help with LVM on Debian 3.01 for S/390
On Wed, Jun 11, 2003 at 07:43:19AM -0600, Hodge, Robert L wrote: I've been trying to define a Logical Volume group on Debian 3.01 for the last week with no success. The current problem is the following error out of vgcreate: It should work fine if you partition the device with fdasd and then pvcreate the partition rather than the whole device. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=162479 -- - mdz
Re: VIPA drivin' me bonkers!
On Tue, Jun 10, 2003 at 05:40:22AM +0800, John Summerfield wrote: I feel impelled to mention that this form is really useful. I often write sed programs thus as I find it more readable than anything else Ive come up with: mv /etc/vim/vimrc /etc/vim/vimrc~ sed /etc/vim/vimrc~ /etc/vim/vimrc \ [...] Wouldn't it be easier to create a ~/.vimrc with the commands you want? -- - mdz
Re: pgp vs gpg
On Wed, Jun 04, 2003 at 11:33:25AM -0500, Tom Duerbusch wrote: My impression is that gpg is basically the same as pgp with the exception that gpg is fully in the public domain. GnuPG is not in the public domain, it is copyrighted by the Free Software Foundation and licensed under the GPL. Great intent... But, just how compatable are they? I haven't found any statement in this area. Where did you look? Question 1.2 in the GnuPG FAQ says: 1.2) Is GnuPG compatible with PGP? In general, yes. GnuPG and newer PGP releases should be implementing the OpenPGP standard. But there are some interoperability problems. See question 5.1 for details. -- - mdz
Re: Linux-hosted turnkey (Re: Turnkey?)
On Tue, Apr 01, 2003 at 09:30:37PM +0800, John Summerfield wrote: 3. Possibly an additional tarball to run Hercules in a chroot environment. Testing changed glibc recently, and that has caused much confusion, more to come. Why would you want to run Hercules in a chroot environment with this type of turnkey CD? -- - mdz
Re: Linux-hosted turnkey (Re: Turnkey?)
On Tue, Apr 01, 2003 at 03:09:30PM +0100, Alan Cox wrote: On Tue, 2003-04-01 at 14:58, Matt Zimmerman wrote: Why would you want to run Hercules in a chroot environment with this type of turnkey CD? Its great for firewall boxes, although not normally needed. I run a mail server on a hercules emulated S/390 running as a nobody user chrooted on a real box. Somehow it feels a bit safer that way That's logical for a network service, but we're talking about something which basically amounts to a software demo. And since you're already giving the software on the CD complete control over the system, a chroot environment within that doesn't give the user any protection. -- - mdz
Re: Linux-hosted turnkey (Re: Turnkey?)
On Wed, Apr 02, 2003 at 01:35:01AM +0800, John Summerfield wrote: My main reason to run stuff in a chroot environment is to ensure I get the right versions of libraries. You could make a tarball on Sarge, and I could unpack it on, lets for for something old, say RHL 3.0.3, and be reasonably confident it would run. Ah, sounds like we are talking about somewhat different projects then...with a live CD setup such as Knoppix, you have complete control over all software installed, so there are no concerns about library compatibility. -- - mdz
Re: Linux-hosted turnkey (Re: Turnkey?)
On Tue, Apr 01, 2003 at 12:55:04PM -0500, John R. Campbell wrote: But there's no real reason to make it a linux-booting CD; Sure there is. It is much easier to provide a complete user experience with this kind of control over the system, and there are many fewer things that can go wrong. If the goal is to provide something which works everywhere, without conflicts with any existing software, requiring zero support, it is the way to go. It doesn't even matter what operating system (if any!) the user has installed; they will always get the same experience. If DVD images (being 4.7GB in size) can be tolerated given the number of such drives becoming standard then a much larger set of installed images becomes possible. With bzip2 compression, it should be possible to fit a useful set of demo images on a normal CD-ROM. So, we're talking a CD that has the Hercules emulator on it (with an x86 linux tree as well) made to autostart in the Windows environment with all the neat-and-keen tools that the MVS 3.8 Tur(n)Key used to show the way. Why would you have an x86 linux tree for a CD intended to autostart in a Windows environment? -- - mdz
Linux-hosted turnkey (Re: Turnkey?)
On Mon, Mar 31, 2003 at 12:03:05PM -0500, Post, Mark K wrote: Rob van der Heij was tossing this idea around last year. He/we didn't get very far on it (so far as I know). One of the things he wanted to do was have all the major Linux/390 flavors on one CD, and that was pushing things just a little bit too much in terms of available space. That sounds like it should fit, actually. My Debian image (bz2 compressed CCKD) is only 54M, and there are not really that many Linux/390 distributions out there. It would be relatively easy to create a Knoppix CD with the Hercules deb installed, include a bunch of DASD images, and add a few icons to the KDE desktop to launch them. No installation required; just boot the CD and click the mouse. I don't have the time or inclination to work on this, but would be willing to provide assistance with the Debian side of things if someone else does. -- - mdz
Re: Linux-hosted turnkey (Re: Turnkey?)
On Mon, Mar 31, 2003 at 06:42:42PM -0800, Ken Dreger wrote: Sounds like this may be fun, I have some extra time and can contribute, where do we start ?? If you haven't already, download a Knoppix ISO, burn a CD, and boot it on a spare machine (desktop, laptop, whatever). It's very neat stuff in its own right, and provides a very full-featured system which runs almost anywhere. There is some information about customization here: http://www.knoppix.net/docs/index.php/FaqCustomising with step-by-step instructions for how to make some changes and build a new ISO image. In our case, you will want to at least: - install the hercules package (the version from Debian unstable if possible, otherwise build from source) - copy in some DASD images (I've already made a Debian one; there may be others out there for other distributions which allow that sort of thing) - write some scripts to start hercules with different configurations, maybe with a shadow file to keep the bulk of the image on the CD and write changes to the ramdisk - add some icons to the KDE desktop to run the scripts This would be a very handy thing to be able to hand out to folks who are interested in experimenting with Linux/390, given the extremely low barrier to entry. -- - mdz
Re: Tomcat RPMs for S390
On Wed, Mar 12, 2003 at 10:09:30AM -0500, Post, Mark K wrote: Hmm. If it really is 100% Java, then it might very well work. I know Apache does, so 50% of the problem is already solved. tomcat has its own web server built in, so in the event that Apache were a problem, it would not need to be solved. :-) -- - mdz
Re: dasd.c fix?
On Thu, Mar 13, 2003 at 03:06:51PM -0600, Stephen Frazier wrote: Is there a way to determine if the fix is installed on my Linux? I have the Debian Linux that was sent out by Sine Nomine Associates in December 2002. Assuming this means you are using the 2.4.17 kernel that ships with Debian 3.0: It includes the following patches released by IBM for the Linux kernel version 2.4.17: * linux-2.4.17-s390.tar.gz (released on 2002.02.05) * linux-2.4.17-s390-1-lcs.tar.gz (released on 2002.03.04) * linux-2.4.17-s390-2.tar.gz (released on 2002.04.15) * linux-2.4.17-s390-3.tar.gz (released on 2002.06.12) * linux-2.4.17-s390-4-iucv.tar.gz(released on 2002.08.16) Additionally it includes * a patch by Gerhard Tonn [EMAIL PROTECTED] which adds support for a second initrd (needed by the s390 boot-floppies); * the cpint-patch by Neal Ferguson which allows to invoke CP commands from Linux; * the cmsfs-patch by Rick Troth [EMAIL PROTECTED] which enables read only access to CMS disks. (from 'apt-cache show kernel-image-2.4.17-s390') The third line item would be the April code drop for the August 2001 stream (2.4.17) I believe. -- - mdz
Re: Postfix Install problem
On Thu, Mar 13, 2003 at 11:51:13PM +0200, Tzafrir Cohen wrote: The alternatives mechanisms was invented by debian and adopted by Mandrake (do you use their packages? Or is it a later version or RH?) It is aimed at solving file-conflicts between packages. For example: both sendmail and postfix have /usr/sbin/sendmail . Not exactly; it was aimed at allowing the user to choose which of a set of packages (which provide compatible interfaces) should be used to provide that interface on the system. /usr/sbin/sendmail is a good example. While several packages can provide /usr/sbin/sendmail, they also usually conflict with one another because they are configured to listen on the smtp port (not because of file conflicts). -- - mdz
Re: Newbie Questions
On Fri, Mar 07, 2003 at 10:56:40AM -0500, Scott Chapman wrote: 1) My feeling is that SuSE is probably the most popular for s390 at the moment, but I don't have any hard data to back that up. If you just want to play Debian is unencumbered by any issues with a vendor wanting money from you. The potential downfall is that ISVs seem to have certified their products for SuSE or RedHat, not Debian. If these ISVs are smart, they will gravitate toward standards like LSB, and certify their products for use on any LSB-compliant distribution. This will make their products available on a wide range of systems, while at the same time giving them a single target to follow. Some of the more adaptive ISVs are also interested in supporting Debian directly. -- - mdz
Re: API to get fullpath name
On Wed, Mar 05, 2003 at 06:46:55PM -0500, Ferguson, Neale wrote: How do I, within a program, get the full pathname of the program I'm executing? argv[0] will have the command name but not necessarily the full pathname. There may not even be a pathname which refers to the program being executed. -- - mdz
Re: Gnome
On Wed, Mar 05, 2003 at 11:40:21PM +0100, Per Jessen wrote: But knoppix does /just/ that. What can be easier than placing the CD in the tray, hit load, then reboot ? No HDD partitioning, no install, no nothing, but Linux KDE up and running in 5mins flat. Not even Windows can do that. Are you implying that Windows is either fast or simple to install? In my experience, it is neither, and the vast majority of users have never even attempted it. -- - mdz
Re: knoppix (was: Gnome)
On Fri, Mar 07, 2003 at 06:08:13PM +0100, Per Jessen wrote: Absolutely. And since you're looking to do an actual install anyway, knoppix won't do you much good. As a matter of fact, it will. http://www.linuxworld.com/site-stories/2002/1104.barr.html -- - mdz
Re: SLES8 Install Problems
On Thu, Feb 27, 2003 at 10:22:55AM +0100, Stefan Gybas wrote: Noll, Ralph wrote: zvmlinx4:/mnt # mount -t smbfs -o username=rnoll //10.201.18.9/cdrom /mnt/cdrom 3481: session request to 10.201.18.9 failed (Called name not present) 3481: session request to 10 failed (Called name not present) Try using the hostname instead of the IP address. It looks like mount is interpreting 10.201.18.9 as a hostname with domain. You might also need to add this host to /etc/hosts or /etc/lmhosts if it can't be found by the mount command. Another possible mistake might be an incorrect domain or workgroup name in /etc/samba/smb.conf or a case mismatch of the share name. Does smbclient -L hostname work? If this is a fully installed system, the answer is probably apt-get install smbfs. If it is the installer, I do not think this will work. -- - mdz
Re: SLES8 Install Problems
On Thu, Feb 27, 2003 at 10:56:26AM +0200, Sergey Korzhevsky wrote: Noll, did you read the subject? I can't find anything about Debian! So it looks like your kernel hasn't support for smbfs. Check smbfs in /proc/filesystems (what was included in the kernel) an check file /lib/modules/YOUR-KERNEL-VERSION/kernel/fs/smbfs/smbfs.o - not included, but presents as module. If you can't find this, recompile your kernel. The problem is likely not with the kernel. Neither the standard mount program nor the busybox one support mounting smbfs shares; they require a separate mount.smbfs program. I do not think this is present in the Debian installer. This works fine using a fully installed system with the smbfs package installed, but to mount an smbfs filesystem in the installer will require some tweaking. -- - mdz
Re: so correct me if I am wrong
On Fri, Feb 21, 2003 at 11:55:16AM -0500, Thomas David Rivers wrote: My experience has been that cross-compiling is risky, at best. While working to port SAPDB, one of the other people helping was doing cross-compiles, and getting different results than I was. Setting up a cross-compile environment is apparently not easy to do correctly, so I try to avoid it as much as possible. Cross-compiling is risky... but, it can be done. Earlier versions of gcc had problems, but it's getting better (and, of course plugSystems/C/plug gets it right.) The problems are not generally with the compiler/assembler/linker toolchain, but with the packages being built. Most programs have never been cross-compiled, and make assumptions that do not hold true for cross-compilation. For example, that the same compiler used to build executables for the target system can be used to build programs meant to run on the build system (such as utilities used during the build). -- - mdz
Re: so correct me if I am wrong
On Sat, Feb 22, 2003 at 04:18:56PM +0200, Tzafrir Cohen wrote: Sounds fishy to me. But anyway, there are slackware-based distoros, RH-based distros, debian-based distros, but there are no (3rd-party) SuSE-based distros . And remember linux distros come in every posible shape and size. So I figure people don't trust this as a solid base. I don't think that is a fair conclusion. Some distributions lend themselves better to serving as a base for other, specialized distributions, but this does not necessarily have anything to do with their quality and robustness. -- - mdz
Re: mingetty and console output
On Fri, Feb 21, 2003 at 02:21:42PM +0100, Florian La Roche wrote: I have not been good at releasing updated mingetty versions. The one in debian is still the one released in 1996. ;-) Last summer I made a re-release by just applying the patches that have been applied to version shipped with Red Hat Linux and named the result mingetty-1.0. Over the last days I have again looked at the source and adapted it to more current Linux releases. I've forwarded a note to the Debian package maintainer notifying him of the availability of this updated release. After a certain amount of time without a release, it is easy not to notice when something like this happens. The Debian package says that the sources came from: ftp://jurix.jura.uni-sb.de/pub/linux/source/system/daemons/mingetty-0.9.4.tar.gz but that host does not appear to run an FTP server anymore. :-) -- - mdz
Re: debian
On Fri, Feb 21, 2003 at 02:54:00PM -0600, Noll, Ralph wrote: I am trying Debian for s/390 under VM how do I install programs.. under Suse i used Yast.. i need FTP installed on Debian http://www.debian.org/doc/manuals/apt-howto/ -- - mdz
Re: debian
On Fri, Feb 21, 2003 at 04:04:28PM -0600, Noll, Ralph wrote: there is nothing in my sources.list This should have been set up during the installation. Was this a CD-ROM or network installation? Anyway, the HOWTO that I showed you contains, in section 2.1, a working example sources.list. Read the first two chapters; they will explain the basics of working with apt. -- - mdz
Re: debian
On Fri, Feb 21, 2003 at 05:30:23PM -0600, Noll, Ralph wrote: http://http.us.debian.org/debian/stable/main/contrib/non-free no directory like the above debian/stable/main/contrib/non-free You are copying the line incorrectly; try cut-and-pasting it instead. Please take this discussion to [EMAIL PROTECTED], where you will find more help. -- - mdz
Re: ISPF for Linux
On Tue, Feb 18, 2003 at 09:59:28PM -0600, Jay Maynard wrote: EMACS has a lot of wired-in assumptions about how editors are supposed to work that make implementing a true WordStar compatibility mode impossible [...] I'd be astounded if someone were to make an EMACS mode that acted like ISPF, and suspect that it'd be no less complex than writing it from the ground up in C. On the contrary, I'd suspect that it would be less complex than viper-mode (which is itself less than 10k lines of elisp according to sloccount). There is very little in emacs that is wired-in, and this is certainly not impossible. Whether it is worthwhile is another matter. -- - mdz
Re: ISPF for Linux + Other Question
On Wed, Feb 19, 2003 at 09:57:30AM -0600, Eric Bielefeld wrote: Wow, I sure got lots of replies to this. There seems to be a lot of controversy over what the best editor is for Linux. On the IBM Linux Tools for Developers page that someone posted, I counted approximately 53 different Linux editors. Indeed, and the available software is much larger still than what is listed on that page. Debian contains nearly 100 editors, many of them workalikes of classic editors and one another. -- - mdz
Re: ISPF for Linux
On Wed, Feb 19, 2003 at 10:00:02PM -0500, David Boyes wrote: Yes, but the point is that it is not installed by default. The only visual editor installed by default is vi, or some clone of it, so you have to know at least a little vi to get along in situations where everything else isn't there. One still occasionally encounters terminals which are incapable of running vi (the Hercules console for example). The ed utility is the standard text editor. - ed(1) (though not the GNU version) -- - mdz
Re: so correct me if I am wrong
On Thu, Feb 20, 2003 at 04:33:54PM -0600, Noll, Ralph wrote: so correct me if i am wrong if you want to stay current with Linux (either Suse or RedHat) you are going to have to pay. if you want to go totally free and do it all yourself you can go Debian yes?? no??? You seem to imply either that Debian does not allow you to stay current, or that it is not Linux. Neither of these is true; why would you think so? -- - mdz
Re: ISPF for Linux
On Tue, Feb 18, 2003 at 11:32:25AM -0600, Eric Bielefeld wrote: Does anyone know if there is an ISPF product for Linux, preferably a free one I can download? If so, what web site do I go to. A quick search turns up at least a couple of things worth investigating. http://www.wrkgrp.com/unispf/index.html uni-SPF is available on all major commercial UNIX platforms. It includes 90 days free support, and support is available on an annual contract basis thereafter. Support entitles you to telephone and email assistance in installing and using the product, problem resolution services, and product upgrades. (doesn't mention Linux, but it'd be foolish to ignore Linux while supporting commercial UNIX implementations) The name SPF-UX also appears, but I don't see any concrete information about it. I'm surprised not to find an Emacs mode for emulating this environment, since the feature sets seem to have a lot of overlap. In terms of what is available as free software (and likely included with your Linux distribution), most resources point to THE (a name that I've always dismissed as impossible to search for). This resource from IBM is worth a look as well: http://www-1.ibm.com/servers/eserver/zseries/os/linux/ldt/slate_dev.html -- - mdz
Re: Installing lvm2
On Thu, Feb 13, 2003 at 10:52:23AM -0600, Dennis Wicks wrote: I am trying to install lvm2. I downloaded the .deb file but could not get apt-get to install it no matter what I tried. I finally got some results with dpkg -i ... but it said there was a dependency on libc6. I downloaded that. Same difficulties as with lvm2. I seem to be in a deendency loop, and I can't find any method that works. Is there a doc somewhere that starts out 1. Download the .deb file? I would really appreciate some help!! It sounds like you are trying to download a package for Debian 'unstable' or 'testing' and install it on a 'stable' system, while it requires newer libraries from 'unstable' or 'testing'. This is doomed to fail without additional tweaking; you must either recompile the source package for stable, or upgrade the dependent packages to testing or unstable. The former is usually the better choice. -- - mdz
Re: Where are dev addresses hiding?
On Wed, Feb 12, 2003 at 12:04:12PM -0600, Dennis Wicks wrote: My query is, where are these addresses hidden so I can change them in my next clone? Or, alternatively, how can I make linux re-sense and re-configure at boot time so I don't have to mess with it? It appears to do this for dasd. /etc/modutils/ctc.chandev After editing this file, run update-modules(8) which will regenerate /etc/chandev.conf. -- - mdz
Re: LVM LVM2
On Mon, Feb 10, 2003 at 12:47:28PM -0600, Dennis Wicks wrote: I am running Debian 2.4.19 under z/VM 3.1.0, and I need some file systems and files that won't fit on a 3390-3. Maybe not on three or four of them! Is LVM2 ready for prime time yet? You would probably get a more complete answer on the LVM2 list. If not, will it be compatible with LVM? From some discussions I have read, I think not, but I have not worked much with LVM2, so take this with a grain of salt. Any hints, tips, caveats or doc pointers I need to know before I plunge into this? Another alternative is EVMS, which is compatible with LVM but also integrates management of RAID volumes, disk partitioning, and other features from a unified interface. EVMS, like LVM2, requires a patch to the 2.4 kernel in order to work. In future stable kernels, both will use the standard device-mapper interface and should not require a patch. Plain LVM and RAID-linear are both available in stock 2.4 kernels for this purpose. -- - mdz
Re: raid question
On Mon, Feb 10, 2003 at 12:57:01PM -0600, Dennis Wicks wrote: I was told that if you lose one volume of a raid set you can replace and the data will get rebuilt, but if you lose a second volume before you get the first one rebuilt you will lose all the data -- irretrievably! If you are talking about RAID-5, and if by rebuilt, you mean the process of writing new redundant data onto a replacement disk, then yes, this is true. However, this does not require the physical replacement of a disk. You can configure additional disks in a RAID-5 volume as spares, which will immediately be used to replace a failed drive without operator intervention. If you require more redundancy than this, you can use RAID-1 (mirroring), which can be extended to any number of disks, of which all but one can fail without losing any data (of course, all data must be written to every disk at once, so performance will suffer). -- - mdz
Re: raid question
On Mon, Feb 10, 2003 at 11:35:45AM -0800, Fargusson.Alan wrote: There is no alternative to RAID. If you have any ideas I would like to here them. Note that RAID protects you from the failure of one drive. Actually one drive for each RAID array, and you can have many arrays. RAID does not protect you from file system corruption and the like. You still need to back up your data. RAID collectively refers to several different methods (RAID levels) of organizing multiple disks into a single storage volume. You can choose the level of failure that the volume will be able to withstand, with corresponding tradeoffs in performance, manageability, and expense. The level that you are talking about here is RAID-5. You can even layer RAID levels together, for example RAID-0 over RAID-1, providing a lot of flexibility. What is the situation that you have where you are looking for an alternative to RAID? -- - mdz
Re: LVM LVM2
On Tue, Feb 11, 2003 at 11:53:01PM +0800, John Summerfield wrote: I thought that EVMS got the flick in favour of LVM2, and that the LVM2 team had decided to develop its management tools for LVM2. OTOH, I could be confusing myself, LVM is not something I've needed to play with yet. LVM2 and EVMS are entirely different projects with similar goals; they are neither mutually exclusive nor interdependent. In its first major release, EVMS had implemented its own kernel runtime component. It was this particular patch which was not accepted into the mainstream kernel, and a lot of unjustified FUD has been spread about the EVMS project as a result. The second major release of EVMS (currently in alpha) is based on the device-mapper kernel component, which provides many of the kernel services that a volume manager needs. LVM2 uses this same kernel component with different userland tools. Here is an announcement from the EVMS team explaining the situation in depth: http://www.uwsg.iu.edu/hypermail/linux/kernel/0211.0/1562.html -- - mdz
Re: patching patch
On Tue, Feb 11, 2003 at 11:40:59AM -0600, Rick Troth wrote: Can someone suggest the best patches against 'patch' to get it compiled cleanly on zSeries? I know it can be done, but I'm missing something. Best I've found thus far is from Debian, but that still doesn't work right. (I consistently get a seek() error when using the resultant executable.) What did SuSE and RH do? What did ThinkBlue do? Heck, what did Marist do? On what distribution are you trying to compile it? The Debian source package compiled on Debian/390 definitely works. Why do you need to rebuild it rather than using the package? What is your test input? Does it work with patch(1) from a vendor package? -- - mdz
Re: Debian install problem
On Wed, Feb 05, 2003 at 02:39:40PM +0100, Maciej Ksiezycki wrote: Is it so because I didn't load the oco.bin file ? I tried to find it on the net but it was nowhere to be found.My distribution doesn't have it either... In the Debian CVS repository, there is a script by Stefan Gybas which will create the second initrd for you: http://cvs.debian.org/*checkout*/boot-floppies/s390-specials/mkinitrd2.sh?rev=HEADonly_with_tag=MAINcontent-type=text/x-sh There are some instructions at the top of the script. It must be run as root on a Linux system, and you must obtain the OCO tar archives from IBM. -- - mdz
Re: Help needed with Debian/390 under hercules configuration errors
On Wed, Jan 29, 2003 at 01:34:25PM +1000, Vic Cross wrote: On 28.01.2003 at 16:18:43, Matt Zimmerman [EMAIL PROTECTED] wrote: When initial configuration runs (when you telnet into it), http://192.168.0.1:3128 will be given as the _default_ proxy address, but you are prompted to change it and it is not used for anything until after this point. \Enter http proxy information, or leave blank for none:\ Does this mean perhaps that the script is not correctly blanking out the proxy information if blank is entered? I do not have an image nearby, so I cannot try it right now, but in these scripts I thought a blank entry (hitting Return with no entry) meant accept the default with no change... What it does is to fill in the box with http://192.168.0.1:3128 as the default, so that if the user simply presses Enter, they get that value. In order to use no proxy, they must delete this string from the text box. This seems to be the only setting which presents such a problem, so perhaps I'll put up an updated image which changes this default. -- - mdz
Re: DASD not recognized in Debian
On Mon, Jan 27, 2003 at 11:02:38AM -0500, Adam Thornton wrote: Debian uses the devfs. Its DASD is named things like /dev/dasd/device/0151/part1 rather than /dev/dasda /dev/dasdb, etc. I'm confused as to how you did this, because Debian can write a correct IPL record for a devfs system. Possibly because he was booted from a non-devfs system at the time it was written. -- - mdz
Re: SuSE Linux - TOC
On Tue, Jan 28, 2003 at 06:53:52AM +0800, John Summerfield wrote: Take a look at rsync for backups. There are scripts around that allow you to maintain more-or-less any number of full backups on disk for instant recovery. One such solution is rdiff-backup (packaged in Debian). Description: Backup program to use deltas for history rdiff-backup is a script that backs up one directory to another. The target directory ends up a copy of the source directory, but extra reverse diffs are stored in a special subdirectory of that target directory, so you can still recover files lost some time ago. The idea is to combine the best features of a mirror and an incremental backup. rdiff-backup also preserves subdirectories, symlinks, special files, permissions, uid/gid ownership (if it is running as root), and modification times. Finally, rdiff-backup can operate in a bandwidth efficient manner over a pipe, like rsync. Thus you can use rdiff-backup and ssh to securely back a hard drive up to a remote location, and only the differences will be transmitted. -- - mdz
Re: SuSE Linux - TOC
On Tue, Jan 28, 2003 at 12:56:44PM +0800, John Summerfield wrote: On Tue, 28 Jan 2003 11:58, you wrote: [rdiff-backup] The rsync solutions differ in that you could offer them as Windows shares via Samba, and users could peruse directories and files as they were at the time of the backup. The rsync solutions are similar in that they're bandwidth-efficient, and you can (don't have to though) use ssh for security. No messing round with patch, all users see is a set of full backups. The volume of disk space required is the sum of one copy of all files backed up plus whatever metadata Linux (or whatever *x you use) stores. Right. The difference is between storing incremental changes to files, and storing complete copies of every modified file. It is a tradeoff between convenience and space efficiency (for large files, a simple rsync backup is very space-inefficient). It's ideal for retrieving that file some twit deleted, recovering fom unintended updates (including viruses), and requires no special action (once setup) by admin types. If that is desirable, this is (in my experience) most conveniently achieved by providing a volume manager snapshot (with, e.g., LVM or EVMS) on a nightly basis, providing an exact, complete filesystem image from some point in the past, which only takes up space for the chunks which have been modified since then. And, it doesn't matter what format the files are. This is true for all of the solutions we have discussed. -- - mdz
Re: Help needed with Debian/390 under hercules configuration errors
On Sat, Jan 18, 2003 at 07:47:52PM -0500, Matt Zimmerman wrote: On Thu, Jan 16, 2003 at 10:53:27PM -0500, Peter J. Farley III wrote: 0% [Connecting to 192.168.0.1 (198.168.0.1)] Err http://security.debian.org stable/updates/main Packages Could not connect to 192.168.0.1:3128 (192.168.0.1), connection timed out ... Failed to fetch http://security.debian.org/dists/stable/updates/main/binary-s390/Packag es Could not connect to 192.168.0.1:3128 (192.168.0.1), connection timed out 192.168.0.1:3128 happens to be where my Squid proxy is. Did that bit of configuration leak into the image that I built somewhere? I'm away from home right now and can't check it, but it should be simple to fix when I get back if this is the case. I have returned home and had a chance to look into this, and upon inspection it appears that the image is fine. When initial configuration runs (when you telnet into it), http://192.168.0.1:3128 will be given as the _default_ proxy address, but you are prompted to change it and it is not used for anything until after this point. Enter http proxy information, or leave blank for none: -- - mdz
Re: Help needed with Debian/390 under hercules configuration errors
On Thu, Jan 16, 2003 at 10:53:27PM -0500, Peter J. Farley III wrote: 0% [Connecting to 192.168.0.1 (198.168.0.1)] Err http://security.debian.org stable/updates/main Packages Could not connect to 192.168.0.1:3128 (192.168.0.1), connection timed out ... Failed to fetch http://security.debian.org/dists/stable/updates/main/binary-s390/Packag es Could not connect to 192.168.0.1:3128 (192.168.0.1), connection timed out 192.168.0.1:3128 happens to be where my Squid proxy is. Did that bit of configuration leak into the image that I built somewhere? I'm away from home right now and can't check it, but it should be simple to fix when I get back if this is the case. -- - mdz
Re: I got unsubscribed...
On Tue, Jan 14, 2003 at 08:02:46AM -0600, Jay Maynard wrote: ...because my mailer rejects any messages with a character set of BIG5, EUC-KR, or KS_C_5601-1987. I got three of those in the past few days. The right thing to do is to accept these messages and discard them, rather than returning an error. That way you don't get unsubscribed from anything, and you also don't generate a lot of unhelpful additional traffic for postmasters around the world (spam generally does not have a useful return address). -- - mdz
