Re: [DNG] lost ability to execute
On Sat, Jun 03, 2017 at 09:00:04PM -0400, Haines Brown wrote: > Ged, thank you for the reply. But it leaves questions in my mind. > > On Sat, Jun 03, 2017 at 02:04:05PM +0100, G.W. Haywood wrote: > > > >Installation of Jessie 1.0.1 AMD64 went well. > > > > Considering what you write below, I'm not entirely convinced. :) > > > > >I did not install desktop packages during install, but instead xorg > > >afterwards along with a lot of other packages. > > > > This starts to look like Pilot Error, but in fairness I suspect that > > what you're trying to do might also be beyond the capacity of the > > system that you're trying to do it with at this stage in its life. > > Devuan Jessie is, in software terms, rather new. If you stress the > > package installation you can expect to break it in interesting ways, > > which may make recovery difficult if you're not an expert. > > I have an installation routine I've used since Lenny and perhaps even > Etch. I stuck to it, and generally accepts defaults. > > The system is a bit old (3 yrs), but has 32 G RAM and 1 T disks. It > should be able to endure an install. > Hi Haines, I don't know exactly what your custom "installation routine" entails, but since there have been a few dozen thousands successful Devuan installations in the last few months, and nobody has reported anything even remotely similar to the problems you talk about, shouldn't we probably consider at least as a remote option the possibility that there is something not entirely right in the custom procedure you are using, or something peculiar in your specific system configuration, that might have little to do with Devuan? HTH KatolaZ -- [ ~.,_ Enzo Nicosia aka KatolaZ - GLUGCT -- Freaknet Medialab ] [ "+. katolaz [at] freaknet.org --- katolaz [at] yahoo.it ] [ @) http://kalos.mine.nu --- Devuan GNU + Linux User ] [ @@) http://maths.qmul.ac.uk/~vnicosia -- GPG: 0B5F062F ] [ (@@@) Twitter: @KatolaZ - skype: katolaz -- github: KatolaZ ] signature.asc Description: Digital signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] lost ability to execute
On Sun, Jun 04, 2017 at 09:54:38AM +0100, KatolaZ wrote: > I don't know exactly what your custom "installation routine" entails, > but since there have been a few dozen thousands successful Devuan > installations in the last few months, and nobody has reported anything > even remotely similar to the problems you talk about, shouldn't we > probably consider at least as a remote option the possibility that > there is something not entirely right in the custom procedure you are > using, or something peculiar in your specific system configuration, > that might have little to do with Devuan? Yes, I agree, but my "routine" for installation is not "custom", but follows installer defaults. I had no problem with the Jessie Beta install, and the first time I installed Jessie 1.0.1 I was able to install many packages before I lost the ability to execute. On subsequent installations I lost ability to execute anything right from start. I could run nano, but not aptititude. Bash found former but not later, although both in /user/bin. It is almost as if I broke hardware when I installed a package. Next time I'll first run mkfs.ext4 on the disk. I put the Jessie amd64 ISO on a key. Because I had incomplete ISO at one point, after doing # dd if=devuan_jessie)1.0.0_amd64_DVD.iso of=/dev/sdd bs=4M; sync I did checksum. I had no trouble booting to it, but could not install the installer components from "CD" because my key not recognized. So used Devuan repository for them instead. Among installer components, I installed the mbr-udeb for PCs because I assumed that meant to use the UBR. But now I wonder if it means to use MBR for 32-bit system. When it came to partitioning, I choose the default no to UEFI. I put /boot into / without breaking it out as I usually do. I don't quite follow the default partition order, and I add two custom partitions. When it came to configure APT, I chose a network mirror. Next time I'll try to get packages from my USB key. Found I had to use http protocol because ftp didn't work. In collections to install, I un-select desktop environment. I suppose the justification for Devuan is to avoid such bloat. After installing GRUB2 the last two times, I go immediately to a shell instead of a reboot. I find that in the /boot/grub/ directory the i386-pc directory is missing, and so was unable to boot the last two installations. This strikes me as a problem of 32/64-bit confusion, but don't know how that is possible. Haines ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] lost ability to execute
Hello Haines, On Sun, 4 Jun 2017, Haines Brown wrote: I have an installation routine I've used since Lenny and perhaps even Etch. I stuck to it, and generally accepts defaults. A list of exactly what you started with and exactly what you did, step-by-step, would be more useful to the problem-finding. You've made a start in your reply to KatolaZ although it's a bit woolly. Don't know what it means to "stress" package installation. What you've done seems to have stressed it. :) It's the same set of packages I had not problem installing on earlier versions of Debian. It looks like the problem (if it's only one, and I'm starting to wonder) happened independently of installing sets of packages. Are you suggesting that Devuan Jessie might have some installation weaknesses not present in Debian Jessie? Because it's different and less well exercised, I'd be prepared to bet on it - but I'm not becessarily sure that that's the problem here. :) How can a package installation damage existing libraries? I'm almost sure that the damage is that you have mixed architectures. You have either a 32-bit binary looking for 32-bit executables or a 64-bit binary looking for 64-bit executables and in neither case is it finding them. When you look in the directory tree you see files which you take to be the executables, but they are not executable by your running system so it perfectly properly ignores them. How is it possible to mix 32-bit and 64-bit if one follows the installer's defaults? You tell us. The list I mentioned above might become the basis of a bug report, or it might be the answer to many of the questions, but at the moment I don't know if there's enough information for anyone to be able to replicate the issue and that's crucial. I installed packages from the US Devuan repository. Perhaps I should instead have installed them from my USB key DVD ISO. I'll try that when I have the time. It could be you've mixed architectures that way but it seems to me more likely that at some point you weren't running the executables that you thought you were running because you had more than one architecture on various operating systems lying around on the drives in your system. > Incidentally mailing list etiquette suggests replies on-list ... My apologies, I didn't see that you were answering my question off list. No, I meant that when I posted to this list you replied privately: Return-Path: Received-SPF: pass (Mechanism 'include:spf.guardedhost.com' matched) mail6: (.net: Sender is authorized to use 'hai...@.net' in 'mfrom' identity (mechanism 'include:spf.guardedhost.com' matched)) Received: from outbound-relay2.guardedhost.com (outbound-relay2.guardedhost.com [IPv6:2607:fe90:1:1::202]) by mail6.jubileegroup.co.uk (8.14.9/8.14.5) with ESMTP id v51IaZ2S018914 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO) for ; Thu, 1 Jun 2017 19:36:45 +0100 In that reply you said While GRUB boots the amd64 Devuan Jessie, its menu does not list my old Debian 32-bit drive. I'll have to pursue this to figure out why. I could do it manually, but there must be an easier way. which information might have been very useful to other list readers. Of course it isn't a 32-bit drive, it's a drive which has installed on it a Debian OS with a 32-bit architecture. It looks suspiciously like you have created some horrible mongrel installation. I think you need to do a clean install of Devuan on a system with a freshly-formatted disc which contains no other OS to confuse the issues. For any fresh install I would tend to remove or disable any drives not necessary to the immediate installation, in case something like GRUB or LILO did something unexpected with them. That may be what's happened here. -- 73, Ged. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] lost ability to execute
On Sun, Jun 04, 2017 at 02:54:22PM +0100, G.W. Haywood wrote: > Hello Haines, > > On Sun, 4 Jun 2017, Haines Brown wrote: > > >I have an installation routine I've used since Lenny and perhaps > >even Etch. I stuck to it, and generally accepts defaults. > > A list of exactly what you started with and exactly what you did, > step-by-step, would be more useful to the problem-finding. You've > made a start in your reply to KatolaZ although it's a bit woolly. I will do that. > It looks like the problem (if it's only one, and I'm starting to > wonder) happened independently of installing sets of packages. Yes, I've come to that conclusion as well. > I'm almost sure that the damage is that you have mixed architectures. > You have either a 32-bit binary looking for 32-bit executables or a > 64-bit binary looking for 64-bit executables and in neither case is it > finding them. When you look in the directory tree you see files which > you take to be the executables, but they are not executable by your > running system so it perfectly properly ignores them. > > >How is it possible to mix 32-bit and 64-bit if one follows the > >installer's defaults? > > You tell us. The list I mentioned above might become the basis of a > bug report, or it might be the answer to many of the questions, but at > the moment I don't know if there's enough information for anyone to be > able to replicate the issue and that's crucial. I have two disks, one with an old Debian Wheezy, my working system, and a new one on which I am trying to install Jessie 1.0.1. To facilitate setup, on my old debian I had created a set of mount points /mnt/devuan/... which fstab automatically mounts when I boot the old debian. I wonder if this could somehow result in a mongrel 32-bit/64-bit installation. When I try a new re-install of Devuan Jessie 1.0.1 later and record exactly what I do, I'll first umount all these mount points and comment their mounts in fstab. Another anomaly is what when I boot the old devuan, the boot goes to recovery mode despite what is selected in the GRUB menu. A control-D continues the boot normally. It showed up only fairly recently. But since I seldom reboot it, I didn't worry much about the issue. The two disks are almost the same, one being a bit older than the other. My Debian is on /dev/sdb and I'm installing to /dev/sda, I wonder whether the mongrolized install might be not be some kind of cross over between the two devices caused by BIOS. > >I installed packages from the US Devuan repository. Perhaps I should > >instead have installed them from my USB key DVD ISO. I'll try that > >when I have the time. > > It could be you've mixed architectures that way but it seems to me more > likely that at some point you weren't running the executables that you > thought you were running because you had more than one architecture on > various operating systems lying around on the drives in your system. Yes, that's the case. The problem is to figure out how they might influence each other. Haines ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] lost ability to execute
On Sun, Jun 04, 2017 at 10:42:23AM -0400, Haines Brown wrote: > Another anomaly is what when I boot the old devuan, the boot goes to > recovery mode despite what is selected in the GRUB menu. A control-D > continues the boot normally. It showed up only fairly recently. But > since I seldom reboot it, I didn't worry much about the issue. The two > disks are almost the same, one being a bit older than the other. My > Debian is on /dev/sdb and I'm installing to /dev/sda, I wonder whether > the mongrolized install might be not be some kind of cross over between > the two devices caused by BIOS. One trick I've used is to place an identification file in the root of each file system. Such as /I-am-old-devian in /dev/sdb and /I-am-new0devuan in /dev/sda. It might help tell them apart independent of BIOS renumbering. Or maybe you're booting a kernel from the old system with a root file system from the new one? I once had two Debians on my system, a stable one and a testing one, and I messed up my /etc/fstab on one of them so i had a root partition from one mounted with a /usr from another. Total screwup when I did a routine security upgrade. Hard to figure out, because a lot of packages worked with mixed system versions. -- hendrik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] lost ability to execute
Did a re-install from DVD ISO on key. I first commented all commands to boot Devuan partitions from my Debian fstab. In BIOS I changed disk priority so that the Debian disk would boot. I booted the Devuan installer and choose expert install. I select all defaults unless otherwise noted. Detect CD drive does see my key with its Jessie 1.0.1 amd64 ISO. I do not choose any installer components. I had assigned the same machine name for both disks, but now change the name for the target disk. In partitioning, I realize that I have a backup USB drive connected, and so disconnect it at this point. Next time I'll remember to have it disconnected, but doubt it makes any difference. The backup disk does not hold a copy of the root partition. I install base system defaults. I configure the package manager. This time I have no problem accessing the DVD ISO for packages. I deselect Desktop Environment for installation. I install GRUB on my target disk. I go to a shell and see that I still don't have a i386-pc directory in /boot/grub, and so naturally when I reboot, the boot fails. I'm about to do another installation, but this time don't do it as expert, but follow all defaults, including auto partitioning and Desktop Environment to see what happens. Haines ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] lost ability to execute
Interim report. Without USB back drive installed, I installed the Jessie DVD ISO from key. Do simply installation routine. When it came to partitioning, I accepted a single partition on entire target disk. Oddly my backup USB drive still reported as a possible target even though it is disconnected. I do the installation from the key rather than Devuan repository. I should note that on one prior installation I had installed mbr-pc because I thought it was asking if I wanted to use the MBR. Could this have messed up my disk? When I reboot to the target disk, all I get is a blinking line cursor upper left. Apparently kernel not seen. I try boot Devuan Jessie from my old disk GRUB menu, but is says no such disk. I suspect I need do a hard disk recovery. I suppose mkfs.ext4 on it would be of no help. Haines ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] reportbug default bts
I just ran into a similar issue trying to use reportbug on a computer I installed Devuan 1.0 Stable a week ago today using the Netinstall ISO. Here is what happened when running reportbug for the first time: $ reportbug flashplugin-nonfree Warning: no reportbug configuration found. Proceeding in novice mode. Detected character set: UTF-8 Please change your locale if this is incorrect. Using 'Nate Bargmann ' as your from address. Getting status for flashplugin-nonfree... Will send report to Devuan (per lsb_release). Querying Devuan BTS for reports on flashplugin-nonfree (source)... Unable to connect to Devuan BTS; continue [y|N|?]? y Maintainer for flashplugin-nonfree is 'Bart Martens '. Looking up dependencies of flashplugin-nonfree... Briefly describe the problem (max. 100 characters allowed). This will be the bug email subject, so keep the summary as concise as possible, for example: "fails to send email" or "does not start with -q option specified" (enter Ctrl+c to exit reportbug without reporting a bug). > reportbug: exiting due to user interrupt. Nevermind that my email address is incorrect as I had not yet created a $HOME/.reportbugrc on that computer. I did create $HOME/.reportbugrc and uncommented the line of "bts devuan", set my email, realname, and replyto variables and nothing else and I still get the: Unable to connect to Devuan BTS; continue [y|N|?]? error message, even on this desktop. The BTS doesn't seem to be down as I can access it via the Web. Is there a hardcoded URL that could be incorrect? To recap, I originally got the "Unable to connect" message on a machine that was a clean Devuan Jessie Stable installation. I also receive it on a machine that was migrated over from Debian Jessie. I have confirmed that the reportbug package is at version 6.6.3+devuan1.3 on both computers. - Nate -- "The optimist proclaims that we live in the best of all possible worlds. The pessimist fears this is true." Ham radio, Linux, bikes, and more: http://www.n0nb.us ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
