Re: New working sparc64 netinst image
Hi, [del] >You missed the important part: Manual kernel installation is no longer >necessary, it's just silo that is missing now besides the partman bug. [del] And I can confirm that the automatic kernel selection works for single CPU boxes like a SUN Ultra 60: Sun Ultra 60 UPA/PCI (UltraSPARC-II 450MHz), No Keyboard OpenBoot 3.27, 512 MB memory installed... ~> cat /proc/cpuinfo cpu : TI UltraSparc II (BlackBird) fpu : UltraSparc II integrated FPU pmu : ultra12 prom: OBP 3.27.0 2000/08/23 15:39 type: sun4u ncpus probed: 1 ncpus active: 1 D$ parity tl1 : 0 I$ parity tl1 : 0 Cpu0ClkTck : 1ad29277 cpucaps : flush,stbar,swap,muldiv,v9,mul32,div32,v8plus,vis MMU Type: Spitfire MMU PGSZs : 8K,64K,512K,4MB The installed kernel: ~> ls /boot/vmlinuz* /boot/vmlinuz /boot/vmlinuz-4.3.0-1-sparc64 Which is right, as there's only one CPU :-) The install went exactly as for the other systems, but network and disk were automatically found and configured, as it's PCI based, HTH, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
Hi, >You missed the important part: Manual kernel installation is no longer >necessary, it's just silo that is missing now besides the partman bug. Yeah, it was late yestderday... The only reason why I did the install was to verify this... And then I didn't confirm it... Shame on me. >I actually helped Rod to reinstall raverin (one of the buildds) with >unstable/sparc64 and we verified that the kernel was installed by >base-installer without any manual intervention. I know, I know, just rub it in ;-{ >I will work on fixing the other two issues the following days, I am >just too busy with many other things at the same time ... Don't try too much; see above CU, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
Hi, [del] >The updated base-installer package is now on the FTP servers. Could >someone perform another test installation and verify that the kernel >is now installed properly without any manual work. Yes, installed with preseeding and... >The partman-ext3 problem still needs to be worked around manually >but we can resolve that later with a dedicated sparc64 installation >mirror which will contain a patched version of partman-ext3. ..the -F workaround in partman... >I assume that we're still missing the part in debian-installer which >runs silo-installer to install silo. I haven't checked yet where >that happens, it's the next item on my TODO list. Once I have figure >that out, I will add sparc64 support to that d-i component. .. and a manual silo install, this time on my noisy V240: root@sele:~# uname -a Linux sele 4.3.0-1-sparc64-smp #1 SMP Debian 4.3.3-5 (2016-01-04) sparc64 GNU/Linux root@sele:~# cat /proc/cpuinfo cpu : TI UltraSparc IIIi (Jalapeno) fpu : UltraSparc IIIi integrated FPU pmu : ultra3i prom: OBP 4.22.19 2006/09/06 23:44 type: sun4u ncpus probed: 2 ncpus active: 2 D$ parity tl1 : 0 I$ parity tl1 : 0 cpucaps : flush,stbar,swap,muldiv,v9,ultra3,mul32,div32,v8plus,vis,vis2 Cpu0ClkTck : 5995f5c0 Cpu1ClkTck : 5995f5c0 MMU Type: Cheetah+ MMU PGSZs : 8K,64K,512K,4MB State: CPU0: online CPU1: online Sorry for the late reply, CU, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 01/13/2016 10:10 PM, Ulrich Teichert wrote: > (...) > Sorry for the late reply, You missed the important part: Manual kernel installation is no longer necessary, it's just silo that is missing now besides the partman bug. I actually helped Rod to reinstall raverin (one of the buildds) with unstable/sparc64 and we verified that the kernel was installed by base-installer without any manual intervention. I will work on fixing the other two issues the following days, I am just too busy with many other things at the same time ... Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/10/2016 02:32 PM, John Paul Adrian Glaubitz wrote: > Ok, patches have been merged now [1, 2]. We just need to wait for the > d-i team to release a new version of the base-installer package which > will include my changes. Then we should retry the netinst image which > shouldn't need to be updated at all since the new base-installer package > comes from the FTP archives. The updated base-installer package is now on the FTP servers. Could someone perform another test installation and verify that the kernel is now installed properly without any manual work. The partman-ext3 problem still needs to be worked around manually but we can resolve that later with a dedicated sparc64 installation mirror which will contain a patched version of partman-ext3. I assume that we're still missing the part in debian-installer which runs silo-installer to install silo. I haven't checked yet where that happens, it's the next item on my TODO list. Once I have figure that out, I will add sparc64 support to that d-i component. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/10/2016 02:32 PM, John Paul Adrian Glaubitz wrote: > Ok, patches have been merged now [1, 2]. Forgot the links: > [1] http://anonscm.debian.org/cgit/d-i/base-installer.git/commit/?id=c60ef0bb9ce9decb1a7d22cb2ea330d2df782547 > [2] http://anonscm.debian.org/cgit/d-i/base-installer.git/commit/?id=5b417cdbcd71bb8715e92dc640a08c576bd83f14 -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/04/2016 10:30 PM, John Paul Adrian Glaubitz wrote: > On 01/04/2016 10:39 PM, Ulrich Teichert wrote: >> Yupp, with usual workarounds (preseeding, editing partman, manual module >> selections, manual kernel and silo install) I was able to use the image >> for an install (I used the older image, the CD was still in the drive): > > Fixing the kernel and silo installation issue is up next. Currently > discussing with the d-i people in #debian-boot. I hope they're going > to merge my sparc64 patches for base-installer soon. Ok, patches have been merged now [1, 2]. We just need to wait for the d-i team to release a new version of the base-installer package which will include my changes. Then we should retry the netinst image which shouldn't need to be updated at all since the new base-installer package comes from the FTP archives. I will also have to figure out how to make debian-installer use the packages provided on the ISO image which would make it possible to perform the installation with patched and unofficial packages for sparc64. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/05/2016 09:52 PM, Ulrich Teichert wrote: > Basically the same way, but I stopped the silo install early, edited > /etc/silo.conf to: > > root=/dev/sdb1 > partition=1 > timeout=100 > > image=/boot/vmlinuz-4.3.0-1-sparc64-smp > label=Linux > initrd=/boot/initrd.img-4.3.0-1-sparc64-smp > > And ran silo per hand. Ok, that's what I did after I realized the silo.conf was corrupted :). Also, good to know that SILO actually works with a compressed kernel, I thought that wouldn't work due to insufficient memory. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, >Fixing the kernel and silo installation issue is up next. Currently >discussing with the d-i people in #debian-boot. I hope they're going >to merge my sparc64 patches for base-installer soon. > >Btw, how did you install silo manually? I did that today with: > >chroot /target >apt-get install linux-image-sparc64 ># adding debian unreleased to /etc/apt/sources.list >apt-get update && apt-get install silo ># this installs silo and makes it bootable, but silo points to ># the wrong kernel by default Basically the same way, but I stopped the silo install early, edited /etc/silo.conf to: root=/dev/sdb1 partition=1 timeout=100 image=/boot/vmlinuz-4.3.0-1-sparc64-smp label=Linux initrd=/boot/initrd.img-4.3.0-1-sparc64-smp And ran silo per hand. >Or did you run silo outside the chroot? No, but that should be possible too, you just have to point it to the right config file. HTH, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 01/04/2016 10:39 PM, Ulrich Teichert wrote: > Yupp, with usual workarounds (preseeding, editing partman, manual module > selections, manual kernel and silo install) I was able to use the image > for an install (I used the older image, the CD was still in the drive): Fixing the kernel and silo installation issue is up next. Currently discussing with the d-i people in #debian-boot. I hope they're going to merge my sparc64 patches for base-installer soon. Btw, how did you install silo manually? I did that today with: chroot /target apt-get install linux-image-sparc64 # adding debian unreleased to /etc/apt/sources.list apt-get update && apt-get install silo # this installs silo and makes it bootable, but silo points to # the wrong kernel by default Or did you run silo outside the chroot? Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, >Problem has been resolved now and systemd is up-to-date again on >sparc64. You will soon be able to use the installer images again >once the mirrors have made their next dinstall run (is run every >6 hours, 0:00 UTC, 6:00 UTC, 12:00 UTC and 18:00 UTC). > >Let me know once you have tested the new netinst image but again, >you might have to wait until after 12:00 UTC. Yupp, with usual workarounds (preseeding, editing partman, manual module selections, manual kernel and silo install) I was able to use the image for an install (I used the older image, the CD was still in the drive): ceteol:~> uname -a Linux ceteol 4.3.0-1-sparc64-smp #1 SMP Debian 4.3.3-2 (2015-12-17) sparc64 GNU/Linux ceteol:~> cat /proc/cpuinfo cpu : TI UltraSparc II (BlackBird) fpu : UltraSparc II integrated FPU pmu : ultra12 prom: OBP 3.25.0 1999/12/03 11:35 type: sun4u ncpus probed: 2 ncpus active: 2 D$ parity tl1 : 0 I$ parity tl1 : 0 cpucaps : flush,stbar,swap,muldiv,v9,mul32,div32,v8plus,vis Cpu0ClkTck : 11a45c3b Cpu1ClkTck : 11a45c3b MMU Type: Spitfire MMU PGSZs : 8K,64K,512K,4MB State: CPU0: online CPU1: online Well done, CU, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 01/02/2016 07:55 PM, John Paul Adrian Glaubitz wrote: >> Jan 2 18:15:07 debootstrap: dpkg: regarding .../ifupdown_0.8.5_sparc64.deb >> containing ifupdown: >> Jan 2 18:15:07 debootstrap: ifupdown breaks systemd (<< 228-3~) >> Jan 2 18:15:07 debootstrap: systemd (version 228-2) is present and >> triggered. >> >> Does somebody know a different way forward? > > Well, the error message is pretty clear. ifupdown wants system 228-3 > or newer while sparc64 has still 228-2 which is a result of the > systemd package not being built on sparc64 yet (it's still building, > see [1]). Problem has been resolved now and systemd is up-to-date again on sparc64. You will soon be able to use the installer images again once the mirrors have made their next dinstall run (is run every 6 hours, 0:00 UTC, 6:00 UTC, 12:00 UTC and 18:00 UTC). Let me know once you have tested the new netinst image but again, you might have to wait until after 12:00 UTC. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/02/2016 07:50 PM, Ulrich Teichert wrote: > Booted up OK (with preseeding as usually), but still lacks the change for > the partitioner. Are these changes part of the packages which are retrieved > over the net? I'm not sure. I was actually sure that debian-installer would use the version of partman-ext3 which I provided with "localudebs". It might be true that the updated package has to be on the servers actually. > Perhaps they haven't been updated on debian-ports? Well, I cannot just build and upload a fixed package to unstable because the original maintainer would probably rather upset. I can, however, upload the package to "unreleased" to see whether that helps. The proper version of the package should be 85+sparc64 which includes the fix. I'll just upload the package to unreleased now, just a second. > Unfortunatly, I'm now running into a new error during install of the base > system: > > Jan 2 18:15:07 debootstrap: dpkg: regarding .../ifupdown_0.8.5_sparc64.deb > containing ifupdown: > Jan 2 18:15:07 debootstrap: ifupdown breaks systemd (<< 228-3~) > Jan 2 18:15:07 debootstrap: systemd (version 228-2) is present and > triggered. > > Does somebody know a different way forward? Well, the error message is pretty clear. ifupdown wants system 228-3 or newer while sparc64 has still 228-2 which is a result of the systemd package not being built on sparc64 yet (it's still building, see [1]). Adrian > [1] https://buildd.debian.org/status/package.php?p=systemd=sid -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/02/2016 08:20 PM, Ulrich Teichert wrote: > Yeah, right, now the build of systemd has failed with the infamous segfault > of xsltproc which I wanted to debug on the system where the install has > failed... Yeah, this should really be fixed. I'll check whether I can hack a work around as this issue is starting to annoy me. > I still can't stand systemd, can we please have back sysvinit as default? I hope you are kidding. But in case you are not, the answer is no since this isn't a simple matter of changing defaults and there is a shitton of reverse dependencies that systemd has. Also, sysvinit is old and unmaintained software. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, >Well, the error message is pretty clear. ifupdown wants system 228-3 >or newer while sparc64 has still 228-2 which is a result of the >systemd package not being built on sparc64 yet (it's still building, >see [1]). Yeah, right, now the build of systemd has failed with the infamous segfault of xsltproc which I wanted to debug on the system where the install has failed... I still can't stand systemd, can we please have back sysvinit as default? Groan, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
Hi, >On 01/01/2016 11:55 PM, John Paul Adrian Glaubitz wrote: >> Almost done. It seems I simply did not built the new base-installer and >> partman-ext3 packages. Will be right back with a new image. > >Ok, updated: > >https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso Thanks, I was too tired yesterday... Turned out I had a timeframe of 15 minutes where the install would have been successful, but read on: Booted up OK (with preseeding as usually), but still lacks the change for the partitioner. Are these changes part of the packages which are retrieved over the net? Perhaps they haven't been updated on debian-ports? Only guessing, though, I don't know much about the installer internals... Anyway, after adding the "-F" to the line in 50format_ext3 per hand, I can confirm that the filesystem gets created. Unfortunatly, I'm now running into a new error during install of the base system: Jan 2 18:15:07 debootstrap: dpkg: regarding .../ifupdown_0.8.5_sparc64.deb containing ifupdown: Jan 2 18:15:07 debootstrap: ifupdown breaks systemd (<< 228-3~) Jan 2 18:15:07 debootstrap: systemd (version 228-2) is present and triggered. Jan 2 18:15:07 debootstrap: Jan 2 18:15:07 debootstrap: dpkg: error processing archive /var/cache/apt/archives/ifupdown_0.8.5_sparc64.deb (--unpack): Jan 2 18:15:07 debootstrap: installing ifupdown would break systemd, and Jan 2 18:15:07 debootstrap: deconfiguration is not permitted (--auto-deconfigure might help) The version of ifupdown on d-p has a timestamp of 2016-01-02 00:10, so I wasn't running into it yesterday. Well, I think I'm blocked here, I can't think around this issue without a local mirror with the older version. Does somebody know a different way forward? CU, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 01/02/2016 08:11 PM, John Paul Adrian Glaubitz wrote: > Yeah, this should really be fixed. I'll check whether I can hack a work > around as this issue is starting to annoy me. I just made a binNMU for docbook-xsl with the proposed patch from #765567 [1] which should alleviate this issue. I will then reschedule all affected packages - including systemd - later on. Adrian > [1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=765567 -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/02/2016 10:06 PM, Anatoly Pugachev wrote: > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=809685 Oh, I forgot to mention, please always add: User: debian-sparc@lists.debian.org Usertags: sparc64 X-Debbugs-CC: debian-sparc@lists.debian.org Just did that manually using the bts command. -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On Wed, Dec 30, 2015 at 6:48 PM, John Paul Adrian Glaubitzwrote: > On 12/30/2015 04:36 PM, Anatoly Pugachev wrote: >> where do i report lvm "bus error" currently ? Just checked with my >> installed sparc64 debian, adding lvm2 package, even simple commands >> give "bus error": > > File a bug against the lvm2 package, please. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=809685
systemd again (was: Re: New working sparc64 netinst image)
Hi, >> I still can't stand systemd, can we please have back sysvinit as default? > >I hope you are kidding. But in case you are not, the answer is no since >this isn't a simple matter of changing defaults and there is a shitton >of reverse dependencies that systemd has. Also, sysvinit is old and >unmaintained software. As I'm not a debian developer, my involvement in this matter is minimal. All I can do is either persuading debian to continue to provide an alternative to systemd (weather or not as default, I don't really care) like open-rc or whatever, even sysvinit or change the distro I'm using. But I won't use systemd on my production servers, ever. If that means switching to a BSD flavour or gentoo or anything, fine. For reasons, see [1] for example, but I'm sorry to stirr this all up again on this ML, where all people are tired of the discussion, CU, Uli [1] http://without-systemd.org/wiki/index.php/Arguments_against_systemd -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
Hi, >On 01/01/2016 08:10 PM, John Paul Adrian Glaubitz wrote: >> Oh, nice. I can certainly just change this in d-i as well. > >Updated NETINST ISO containing both changes: > >> >https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso Tried on my Ultra-2, boots OK, but both issues are still there :-( Partitioning is OK, but creating an ext4 filesystems still gets no further than 33%. Looks unchanged to me - from the outside. After manually preparing a filesystem to install on, now different errors are showing up in the log: Jan 1 21:05:12 in-target: (Reading database ... 100%^M Jan 1 21:05:12 in-target: (Reading database ... Jan 1 21:05:12 in-target: 8999 files and directories currently installed.) Jan 1 21:05:12 in-target: ^M Jan 1 21:05:12 in-target: Preparing to unpack .../libc-l10n_2.21-6_all.deb ... Jan 1 21:05:12 in-target: ^M Jan 1 21:05:13 in-target: Unpacking libc-l10n (2.21-6) ... Jan 1 21:05:13 in-target: ^M Jan 1 21:05:15 kernel: [ 1665.647161] dpkg-deb[17759]: segfault at 5c98 ip f801005326c8 (rpc f80100532650) sp 07feff858aa1 error 30001 in libc-2.21.so[f8010047c000+15c000] Jan 1 21:05:15 kernel: [ 1665.647595] BUG: Bad rss-counter state mm:f800238f89e0 idx:1 val:1 Jan 1 21:05:15 kernel: [ 1665.723997] BUG: non-zero nr_ptes on freeing mm: 1 Jan 1 21:05:15 in-target: dpkg-deb: error: subprocess tar was killed by signal --More-- (98% of 1316(Segmentation fault) Jan 1 21:05:15 in-target: ^M Jan 1 21:05:15 in-target: dpkg: error processing archive /var/cache/apt/archives/locales_2.21-6_all.deb (--unpack):^M Jan 1 21:05:15 in-target: subprocess dpkg-deb --control returned error exit status 2^M Jan 1 21:05:15 in-target: Errors were encountered while processing: Jan 1 21:05:15 in-target: ^M Jan 1 21:05:15 in-target: /var/cache/apt/archives/locales_2.21-6_all.deb Jan 1 21:05:15 in-target: ^M Jan 1 21:05:18 in-target: E Jan 1 21:05:18 in-target: : Jan 1 21:05:18 in-target: Sub-process /usr/bin/dpkg returned an error code (1) Jan 1 21:05:18 in-target: Jan 1 21:05:20 localechooser: error: the command 'validlocale' is not available Jan 1 21:05:23 base-installer: error: Unable to execute /usr/lib/post-base-installer.d/60partman-zfs Jan 1 21:05:32 base-installer: info: could not determine kernel flavour Jan 1 21:05:32 base-installer: info: could not determine kernel flavour Jan 1 21:05:32 base-installer: info: could not determine kernel flavour Jan 1 21:05:32 base-installer: info: could not determine kernel flavour Jan 1 21:05:32 base-installer: info: Found kernels '' Jan 1 21:06:41 base-installer: error: exiting on error base-installer/kernel/no--More-- (99% of 1316-kernels-found Jan 1 21:06:44 main-menu[151]: WARNING **: Configuring 'bootstrap-base' failed with error code 1 Jan 1 21:06:44 main-menu[151]: WARNING **: Menu item 'bootstrap-base' failed. Never seen the kernel message in the logs before, HTH, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 01/01/2016 11:55 PM, John Paul Adrian Glaubitz wrote: > Almost done. It seems I simply did not built the new base-installer and > partman-ext3 packages. Will be right back with a new image. Ok, updated: > https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/01/2016 10:35 PM, Ulrich Teichert wrote: > Partitioning is OK, but creating an ext4 filesystems still gets no further > than 33%. Looks unchanged to me - from the outside. Did you check whether the change that Kieron was talking about: > Modified /lib/partman/commit.d/50format_ext3 > and changed the line > mkfs.$filesystem $device $usage >/dev/null; then > to > mkfs.$filesystem -F $device $usage >/dev/null; then was actually in? I might have just made a mistake while creating the new image and forgot to include the updated installer packages. > Never seen the kernel message in the logs before, That's unrelated and probably an issue with your hardware. -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/01/2016 11:48 PM, Ulrich Teichert wrote: > No harm done, though, installed the rest by hand for now and got a booting > system with kernel 4.3.3. That's more than I could say last year ;-) Almost done. It seems I simply did not built the new base-installer and partman-ext3 packages. Will be right back with a new image. -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, >On 01/01/2016 10:35 PM, Ulrich Teichert wrote: >> Partitioning is OK, but creating an ext4 filesystems still gets no further >> than 33%. Looks unchanged to me - from the outside. > >Did you check whether the change that Kieron was talking about: > >> Modified /lib/partman/commit.d/50format_ext3 >> and changed the line >> mkfs.$filesystem $device $usage >/dev/null; then >> to >> mkfs.$filesystem -F $device $usage >/dev/null; then > >was actually in? I might have just made a mistake while creating the >new image and forgot to include the updated installer packages. Definitely not in: if log-output -t partman --pass-stdout \ mkfs.$filesystem $device $usage >/dev/null; then sync status=OK else status=failed fi No harm done, though, installed the rest by hand for now and got a booting system with kernel 4.3.3. That's more than I could say last year ;-) I'll keep it available for further tests, HTH, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
Any chance that we are about ready to write some instructions on how other can setup there own instance of SPARC64 yet? Feeling like my instructions right now would require a bit too much hacking at this point but maybe with the next release of the install iso? On Fri, Jan 1, 2016 at 5:55 PM, John Paul Adrian Glaubitz < glaub...@physik.fu-berlin.de> wrote: > On 01/01/2016 11:48 PM, Ulrich Teichert wrote: > > No harm done, though, installed the rest by hand for now and got a > booting > > system with kernel 4.3.3. That's more than I could say last year ;-) > > Almost done. It seems I simply did not built the new base-installer and > partman-ext3 packages. Will be right back with a new image. > > -- > .''`. John Paul Adrian Glaubitz > : :' : Debian Developer - glaub...@debian.org > `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de > `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 > >
Re: New working sparc64 netinst image
On 01/02/2016 12:08 AM, Kieron Gillespie wrote: > Any chance that we are about ready to write some instructions on how > other can setup there own instance of SPARC64 yet? Actually, once we have resolved these two issues, it should be almost straight-forward to install sparc64. You just need to supply the preseed file at the boot prompt: boot: install preseed/url=http://users.physik.fu-berlin.de/~glaubitz/preseed-sparc64.cfg -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/01/2016 10:49 PM, John Paul Adrian Glaubitz wrote: > was actually in? I might have just made a mistake while creating the > new image and forgot to include the updated installer packages. Working on a new image, just a second. -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/01/2016 08:10 PM, John Paul Adrian Glaubitz wrote: > Oh, nice. I can certainly just change this in d-i as well. Updated NETINST ISO containing both changes: > https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/01/2016 03:13 PM, Ulrich Teichert wrote: > - the installer went ahead downloading, but failed to find a kernel for > installation. Manually installing it in the target via chroot works. Can I get the exact error message, preferably from the log files under /var/log? My suspicion is that debian-installer tries to install the wrong kernel version and therefore fails. Apparently, being able to install the proper kernel version is a bit tricky and according to Helge Deller (CC) is the main reason why he uses snapshotted debian-ports.org mirrors for the installation. I have the suspicion that debian-installer looks for a particular kernel version which was determined at build time. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, and happy new year to all :-) >I've updated the NETINST image [1] now: > >- made some silo improvements for d-i [2] >- disabled GPG verification for the archive in the code I've attempted an install on an Ultra-2 today, auto detection of network hardware and disks failing, but manually selecting the drivers went OK. With preseeding, I was able to select debian-ports as package source. Then it didn't went so well. - partitioning was OK, creating filesystems on the new partitions hangs after 33%. Even on this system, where SUN slices are the right partition scheme for the boot disks. I was able to create the filesystem from an older installation and do the target selection in the new installer afterwards. So, the partitions are setup OK. - the installer went ahead downloading, but failed to find a kernel for installation. Manually installing it in the target via chroot works. That's where I have to stop now, getting hungry :-) HTH, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 01/01/2016 07:45 PM, Ulrich Teichert wrote: > Jan 1 18:15:11 base-installer: error: Unable to execute > /usr/lib/post-base-installer.d/60partman-zfs > Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > Jan 1 18:15:20 base-installer: info: Found kernels '' Aha, base-installer it is. Looking at the code: root@test-adrian1:/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer# cd packages/base-installer/kernel/ root@test-adrian1:/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer/packages/base-installer/kernel# ls alpha.sh arm64.sh armel.sh hppa.sh i386.sh kfreebsd-amd64.sh m68k.sh mipsel.sh powerpc.sh READMEsh4.shtests amd64.sh armeb.sh armhf.sh hurd-i386.sh ia64.sh kfreebsd-i386.sh Makefile mips.shppc64el.sh s390x.sh sparc.sh root@test-adrian1:/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer/packages/base-installer/kernel# There is apparently no sparc64 support in the base-installer package. Will fix that and upload a new ISO image :). Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, >Can I get the exact error message, preferably from the log files under >/var/log? My suspicion is that debian-installer tries to install the >wrong kernel version and therefore fails. > >Apparently, being able to install the proper kernel version is a bit >tricky and according to Helge Deller (CC) is the main reason why he >uses snapshotted debian-ports.org mirrors for the installation. > >I have the suspicion that debian-installer looks for a particular >kernel version which was determined at build time. There are quite a number of interesting entries in the syslog of the installer. Complete log after disk detection: Jan 1 17:54:34 partman: Reading all physical volumes. This may take a while... Jan 1 17:57:50 kernel: [ 1578.055815] Adding 979952k swap on /dev/sda2. Priority:-1 extents:1 across:979952k FS Jan 1 17:57:50 kernel: [ 1578.092955] Adding 224896k swap on /dev/sdb2. Priority:-2 extents:1 across:224896k FS Jan 1 17:57:52 kernel: [ 1579.899054] EXT4-fs (sdb1): mounting ext3 file system using the ext4 subsystem Jan 1 17:57:52 kernel: [ 1579.938840] EXT4-fs (sdb1): warning: checktime reached, running e2fsck is recommended Jan 1 17:57:52 kernel: [ 1579.949878] EXT4-fs (sdb1): mounted filesystem with ordered data mode. Opts: errors=remount-ro Jan 1 17:57:55 apt-install: Queueing package e2fsprogs for later installation Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/init.d/26zfs-device--More-- (31% of 1309s: Jan 1 17:57:56 main-menu[150]: (process:4176): line 42: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/choose_partition/60partition_tree/do_option: Jan 1 17:57:56 main-menu[150]: (process:4176): line 88: Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/active_partition/copy/choices: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/choose_partition/60partition_tree/do_option: Jan 1 17:57:56 main-menu[150]: (process:4176): line 88: Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/active_partition/copy/choices: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/choose_partition/60partition_tree/do_option: Jan 1 17:57:56 main-menu[150]: (process:4176): line 88: Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/active_partition/copy/choices: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/choose_partition/60--More-- (32% of 1309partition_tree/do_option: Jan 1 17:57:56 main-menu[150]: (process:4176): line 88: Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/active_partition/copy/choices: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/choose_partition/60partition_tree/do_option: Jan 1 17:57:56 main-menu[150]: (process:4176): line 88: Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/active_partition/copy/choices: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/fstab.d/zfs: Jan 1 17:57:56 main-menu[150]: (process:4176): line 49: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/fstab.d/zfs: Jan 1 17:57:56 main-menu[150]: (process:4176): line 49: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/fstab.d/zfs: Jan 1 17:57:56 main-menu[150]: (process:4176): line 49: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/fstab.d/zfs: Jan 1 17:57:56 main-menu[150]: (process:4176): line 49: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/fstab.d/zfs: Jan 1 17:57:56 main-menu[150]: (process:4176): line 49: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:57:56 main-menu[150]: (process:4176): /lib/partman/fstab.d/zfs: Jan 1 17:57:56 main-menu[150]: (process:4176): line 49: Jan 1 17:57:56 main-menu[150]: (process:4176): zpool: not found Jan 1 17:57:56 main-menu[150]: (process:4176): Jan 1 17:58:04 main-menu[150]: INFO: Menu item 'bootstrap-base' selected Jan 1 17:58:04 base-installer: warning: Unknown architecture 'sparc64'. Then packages are installed, where some warnings wrt dependecies are
Re: New working sparc64 netinst image
So I was able to get past the 33% error on the partition by setting the -F flag for mkfs.ext4 in the debian install scripts. Guidance to the issue was found here https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=767682 Modified /lib/partman/commit.d/50format_ext3 and changed the line mkfs.$filesystem $device $usage >/dev/null; then to mkfs.$filesystem -F $device $usage >/dev/null; then Not sure if this is the best way to handle this issue but it did allow me to proceed further in testing the install process on my Sun Blade 2500. On Fri, Jan 1, 2016 at 9:16 AM, John Paul Adrian Glaubitz < glaub...@physik.fu-berlin.de> wrote: > On 01/01/2016 03:13 PM, Ulrich Teichert wrote: > > - the installer went ahead downloading, but failed to find a kernel for > > installation. Manually installing it in the target via chroot works. > > Can I get the exact error message, preferably from the log files under > /var/log? My suspicion is that debian-installer tries to install the > wrong kernel version and therefore fails. > > Apparently, being able to install the proper kernel version is a bit > tricky and according to Helge Deller (CC) is the main reason why he > uses snapshotted debian-ports.org mirrors for the installation. > > I have the suspicion that debian-installer looks for a particular > kernel version which was determined at build time. > > Adrian > > -- > .''`. John Paul Adrian Glaubitz > : :' : Debian Developer - glaub...@debian.org > `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de > `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 > >
Re: New working sparc64 netinst image
On 01/01/2016 08:08 PM, Kieron Gillespie wrote: > Guidance to the issue was found here > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=767682 > > Modified /lib/partman/commit.d/50format_ext3 > > and changed the line > > mkfs.$filesystem $device $usage >/dev/null; then > > to > > mkfs.$filesystem -F $device $usage >/dev/null; then Oh, nice. I can certainly just change this in d-i as well. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, >On 01/01/2016 07:45 PM, Ulrich Teichert wrote: >> Jan 1 18:15:11 base-installer: error: Unable to execute >> /usr/lib/post-base-installer.d/60partman-zfs >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour >> Jan 1 18:15:20 base-installer: info: Found kernels '' > >Aha, base-installer it is. Looking at the code: > >root@test-adrian1:/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer# >cd packages/base-installer/kernel/ >root@test-adrian1:/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer/packages/base-installer/kernel# >ls >alpha.sh arm64.sh armel.sh hppa.sh i386.sh kfreebsd-amd64.sh >m68k.sh mipsel.sh powerpc.sh READMEsh4.shtests >amd64.sh armeb.sh armhf.sh hurd-i386.sh ia64.sh kfreebsd-i386.sh >Makefile mips.shppc64el.sh s390x.sh sparc.sh >root@test-adrian1:/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer/packages/base-installer/kernel# > >There is apparently no sparc64 support in the base-installer package. > >Will fix that and upload a new ISO image :). That's cool, prio to that, there was: [del] Jan 1 17:58:04 main-menu[150]: INFO: Menu item 'bootstrap-base' selected Jan 1 17:58:04 base-installer: warning: Unknown architecture 'sparc64'. [del] So that's probably the root cause? HTH, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
Was wondering if there is any way to install the SILO bootloader with this installed? I was pretty much able to get everything installed, with all the work arounds listed here, except for SILO which appears to no longer be available in the repros. On Fri, Jan 1, 2016 at 2:20 PM, Ulrich Teichertwrote: > Hi, > > >On 01/01/2016 07:45 PM, Ulrich Teichert wrote: > >> Jan 1 18:15:11 base-installer: error: Unable to execute > /usr/lib/post-base-installer.d/60partman-zfs > >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > >> Jan 1 18:15:20 base-installer: info: could not determine kernel flavour > >> Jan 1 18:15:20 base-installer: info: Found kernels '' > > > >Aha, base-installer it is. Looking at the code: > > > >root@test-adrian1 > :/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer# > >cd packages/base-installer/kernel/ > >root@test-adrian1 > :/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer/packages/base-installer/kernel# > >ls > >alpha.sh arm64.sh armel.sh hppa.sh i386.sh kfreebsd-amd64.sh > >m68k.sh mipsel.sh powerpc.sh READMEsh4.shtests > >amd64.sh armeb.sh armhf.sh hurd-i386.sh ia64.sh kfreebsd-i386.sh > >Makefile mips.shppc64el.sh s390x.sh sparc.sh > >root@test-adrian1 > :/srv/sid-sparc64-sbuild/debian/DEBIAN_INSTALLER/debian-installer/packages/base-installer/kernel# > > > >There is apparently no sparc64 support in the base-installer package. > > > >Will fix that and upload a new ISO image :). > > That's cool, prio to that, there was: > > [del] > Jan 1 17:58:04 main-menu[150]: INFO: Menu item 'bootstrap-base' selected > Jan 1 17:58:04 base-installer: warning: Unknown architecture 'sparc64'. > [del] > > So that's probably the root cause? > > HTH, > Uli > -- > Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: > Stormweg 24 |Imperial State Electric, The Estranged, > 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu. > >
Re: New working sparc64 netinst image
On 01/01/2016 08:32 PM, Kieron Gillespie wrote: > Was wondering if there is any way to install the SILO bootloader with > this installed? I was pretty much able to get everything installed, with > all the work arounds listed here, except for SILO which appears to no > longer be available in the repros. silo is available in 'unreleased' on ftp.debian-ports.org, so you have to make sure you preseed with the unreleased suite included. Try adding it to the preseed file. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 01/01/2016 08:35 PM, John Paul Adrian Glaubitz wrote: > Try adding it to the preseed file. Like this one: http://users.physik.fu-berlin.de/~glaubitz/preseed-sparc64.cfg -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Hi, sorry for the late update, I had a *long* shopping list to run after yesterday to avoid shopping today :-) >I've updated the NETINST image [1] now: > >- made some silo improvements for d-i [2] >- disabled GPG verification for the archive in the code Doesn't seem to work, the installer still refuses to take debian-ports as a package source: Dec 31 12:31:45 main-menu[176]: INFO: Menu item 'choose-mirror' selected Dec 31 12:31:45 anna-install: Queueing udeb apt-mirror-setup for later installation Dec 31 12:32:42 choose-mirror[906]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/oldstable/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 31 12:32:43 choose-mirror[906]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/stable/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 31 12:32:43 choose-mirror[906]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/testing/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 31 12:32:43 choose-mirror[906]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/uures):' E '^(Suite|Codename|Architect--More-- (96% of 49835 bytes) Dec 31 12:32:44 choose-mirror[906]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/sid/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 31 12:32:45 choose-mirror[906]: INFO: suite/codename set to: unstable/sid Dec 31 12:32:45 choose-mirror[906]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian//dists/sid/main/binary-sparc64/Release -O - | grep ^Architecture: Dec 31 12:32:45 anna-install: Queueing udeb sid-support for later installation Dec 31 12:32:45 main-menu[176]: (process:899): wget: server returned error: HTTP/1.1 404 Not Found Dec 31 12:32:45 main-menu[176]: (process:899): wget: server returned error: HTTP/1.1 404 Not Found Dec 31 12:32:45 main-menu[176]: (process:899): wget: server returned error: HTTP/1.1 404 Not Found Dec 31 12:32:45 main-menu[176]: DEBUG: resolver (libc6-udeb): package doesn't exist (ignored) Dec 31 12:32:45 main-menu[176]: INFO: Menu item 'download-installer' selected Dec 31 12:32:46 net-retriever: gpgv: Dec 31 12:32:46 net-retriever: Signature made Thu Dec 31 07:44:13 2015 UTC using RSA key ID C448326E Dec 31 12:32:46 net-retriever: gpgv: Dec 31 12:32:46 net-retriever: Can't check signature: public key not found Dec 31 12:32:46 net-retriever: error: Bad signature on /tmp/net-retriever-963-Re--More-- (99% of 49835lease. Dec 31 12:33:05 anna[959]: WARNING **: bad d-i Packages file Dec 31 12:33:05 main-menu[176]: INFO: Menu item 'download-installer' succeeded b--More-- (99% of 49835ut requested to be left unconfigured. Dec 31 12:33:05 main-menu[176]: DEBUG: resolver (libc6-udeb): package doesn't ex--More-- (99% of 49835ist (ignored) Prior to that, I manually configured the network after DHCP succeeded, as I like to have static IPs for servers, but the installer doesn't seem to care afterwards and still used the DHCP config? Or am I misinterpreting the logging here (IP addresses XXXed/YYYed/ZZZed out): Dec 31 12:31:41 netcfg[782]: ARPING to 192.168.XXX.YYY from 192.168.XXX.ZZZ via enp0s2f0 Dec 31 12:31:41 netcfg[782]: Unicast reply from 192.168.XXX.YYY [58:6d:8f:a5:8a:f] 0.348ms Dec 31 12:31:41 netcfg[782]: Sent 1 probe(s) (1 broadcast(s)) Dec 31 12:31:41 netcfg[782]: Received 1 replies (0 request(s), 0 broadcast(s)) Dec 31 12:31:41 netcfg[782]: INFO: Gateway reachable on enp0s2f0 Dec 31 12:31:42 netcfg[782]: DEBUG: Getting default hostname from rDNS lookup of static-configured address 192.168.XXX.ZZZ Dec 31 12:31:42 netcfg[782]: DEBUG: Hostname found: sele.nms.ulrich-teichert.org Dec 31 12:31:42 netcfg[782]: DEBUG: sele.nms.ulrich-teichert.org is a valid FQDN Dec 31 12:31:42 netcfg[782]: DEBUG: We have a real FQDN Dec 31 12:31:42 netcfg[782]: DEBUG: Preseeding domain as well: nms.ulrich-teichert.org Dec 31 12:31:45 netcfg[782]: DEBUG: No interface given; clearing /etc/network/interfaces Dec 31 12:31:45 netcfg[782]: DEBUG: Writing informative header Dec 31 12:31:45 netcfg[782]: DEBUG: Success! Dec 31 12:31:45 netcfg[782]: DEBUG: Writing loopback interface Dec 31 12:31:45 netcfg[782]: DEBUG: Success! Dec 31 12:31:45 netcfg[782]: DEBUG: Writing static IPv4 stanza for enp0s2f0 Dec 31 12:31:45 netcfg[782]: INFO: Detected enp0s2f0 as a hotpluggable device Dec 31 12:31:45 netcfg[782]: DEBUG: Success! Dec 31 12:31:45 main-menu[176]: (process:781): udhcpc (v1.22.1) started Dec 31 12:31:45 main-menu[176]: (process:781): Sending discover... Dec 31 12:31:45 main-menu[176]: (process:781): Sending select for 192.168.XXX.PPP... Dec 31 12:31:45 main-menu[176]: (process:781): Lease of 192.168.XXX.PPP obtained, lease time 86400 Dec 31 12:31:45 main-menu[176]: DEBUG: resolver (libc6-udeb): package doesn't exist (ignored) But that may be a generic installer bug, I believe. Anyway, trying preseeding now to get
Re: New working sparc64 netinst image
On 12/31/2015 01:55 PM, Ulrich Teichert wrote: > Doesn't seem to work, the installer still refuses to take debian-ports > as a package source: > (...) Alright, thanks for the feedback. I will do some more debugging next week when I am back in the office and have access to a SPARC machine. > Prior to that, I manually configured the network after DHCP succeeded, as > I like to have static IPs for servers, but the installer doesn't seem > to care afterwards and still used the DHCP config? Or am I misinterpreting > the logging here (IP addresses XXXed/YYYed/ZZZed out): > > But that may be a generic installer bug, I believe. Anyway, trying preseeding > now to get around it, If that's the case it should affect amd64 installs as well and you should file a bug report against the "debian-installer" package. Let me know whether you see any other issues, especially the installation issue with the kernel that Anatoly reported. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/30/2015 04:36 PM, Anatoly Pugachev wrote: > first, i had problems with formatting as well, but reused / copied > disk from oracle sparc linux with gpt disk partition table > (description of this problem in my previous email). This is a potential bug in debian-installer which always asssumes Sun partition tables on sparc64 but I'm not sure. It should allow GPT partition tables, too, in any case. > after solving problem with formatting, d-i did not installed kernel, > and i need to install it manually "apt-get install > linux-image-sparc64-smp" , changing silo.conf and updating bootloader. What exactly was the error message? > where do i report lvm "bus error" currently ? Just checked with my > installed sparc64 debian, adding lvm2 package, even simple commands > give "bus error": File a bug against the lvm2 package, please. > Adrian, yeah, lets make this as buildd, about 200 vcpus , 200Gb ram > and 300Gb hdd avail for this task. I'll send you an email with my public SSH key. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
Oh for the record,.. prtconf,alas, frequently ignores stuff that could be reported in ascii supported-labels: 67707400.73756e00.6d627200 => gpt sun mbr > gpt_avail=iutil.execWithCapture("prtconf", ["-v", "-p"]) > if ( gpt_avail.find("gpt") != -1 ): > _disklabel_types = ["gpt", "sun"] > else: > _disklabel_types = ["sun", "gpt"] > > > Node 0xf025eec4 > *gpt*: > supported-labels: 67707400.73756e00.6d627200 > name: 'disk-label' > > >
Re: New working sparc64 netinst image
Hello! I've updated the NETINST image [1] now: - made some silo improvements for d-i [2] - disabled GPG verification for the archive in the code Cheers, Adrian > [1] https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso > [2] https://lists.debian.org/debian-boot/2015/12/msg00299.html -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 30/12/15 15:48, John Paul Adrian Glaubitz wrote: > On 12/30/2015 04:36 PM, Anatoly Pugachev wrote: >> first, i had problems with formatting as well, but reused / copied >> disk from oracle sparc linux with gpt disk partition table >> (description of this problem in my previous email). > This is a potential bug in debian-installer which always asssumes > Sun partition tables on sparc64 but I'm not sure. It should allow > GPT partition tables, too, in any case. > >> after solving problem with formatting, d-i did not installed kernel, >> and i need to install it manually "apt-get install >> linux-image-sparc64-smp" , changing silo.conf and updating bootloader. > What exactly was the error message? > >> where do i report lvm "bus error" currently ? Just checked with my >> installed sparc64 debian, adding lvm2 package, even simple commands >> give "bus error": > File a bug against the lvm2 package, please. > >> Adrian, yeah, lets make this as buildd, about 200 vcpus , 200Gb ram >> and 300Gb hdd avail for this task. > I'll send you an email with my public SSH key. > > Adrian > mm GPT tables,.. bane of my life at one point 8/ you need to find out of your sparc supports gpt tables in the first place for a system disk, since that is a function of the OBP firmware. (once the kernel is running you can have any partitioning schema you like,.. just NOT on the system boot drive. In L4S (anaconda) I was a bit lazy, during the CD install while its still running from RAM, I made a call out to prtconf and looked for gpt. however that required a few changes to prtconf for it to chuck out a string version gpt_avail=iutil.execWithCapture("prtconf", ["-v", "-p"]) if ( gpt_avail.find("gpt") != -1 ): _disklabel_types = ["gpt", "sun"] else: _disklabel_types = ["sun", "gpt"] Node 0xf025eec4 *gpt*: supported-labels: 67707400.73756e00.6d627200 name: 'disk-label'
Re: New working sparc64 netinst image
On 12/30/2015 03:29 AM, Anatoly Pugachev wrote: > On Tue, Dec 29, 2015 at 3:39 PM, John Paul Adrian Glaubitz >wrote: >> On 12/29/2015 01:31 PM, John Paul Adrian Glaubitz wrote: >>> http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso > > installed from this cd/iso version into LDOM, using local mirror and > preseed configuration Awesome, great to hear! Were there any issues you encountered? - spend too much time actually... Yeah, this is currently inevitable. This is still work-in-progress and as long as I don't have physical hardware for testing, it's hard to make the proper improvements. > now i have fully 64bit unstable/sid with 4.3 kernel and systemd working. Great. Let me know if you want me to set this up as a buildd. Since this machine is extremely fast, we could run several buildd instances without any noticable slow-down. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 30 December 2015 at 10:19, John Paul Adrian Glaubitz < glaub...@physik.fu-berlin.de> wrote: > On 12/30/2015 03:29 AM, Anatoly Pugachev wrote: > > On Tue, Dec 29, 2015 at 3:39 PM, John Paul Adrian Glaubitz > >wrote: > >> On 12/29/2015 01:31 PM, John Paul Adrian Glaubitz wrote: > >>> > http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso > > > > installed from this cd/iso version into LDOM, using local mirror and > > preseed configuration > > Awesome, great to hear! Were there any issues you encountered? > > - spend too much time actually... > > Yeah, this is currently inevitable. This is still work-in-progress and > as long as I don't have physical hardware for testing, it's hard to > make the proper improvements. > > > now i have fully 64bit unstable/sid with 4.3 kernel and systemd working. > > Great. Let me know if you want me to set this up as a buildd. Since > this machine is extremely fast, we could run several buildd instances > without any noticable slow-down. > > Adrian > > -- > .''`. John Paul Adrian Glaubitz > : :' : Debian Developer - glaub...@debian.org > `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de > `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 > > No go for me unfortunately. I'm using a preseed file with only "d-i debian-installer/allow_unauthenticated boolean true". Start the installer with, boot: install preseed/url=http://x-server.olberg.name/preseed-sparc.cfg and as repo ftp.debian-ports.org/debian/ (which seems to work fine). Which seems to work fine, until I get to the formatting part, where it seems to hang at 33%. This is in an LDOM on a T5140, tried changing from zvol to file, also tried ext2 and ext3 instead of ext4. Same result every time. -kp
Re: New working sparc64 netinst image
On 12/30/2015 01:18 PM, Knut Petter Ølberg wrote: > Which seems to work fine, until I get to the formatting part, where it > seems to hang at 33%. This is in an LDOM on a T5140, tried changing from > zvol to file, also tried ext2 and ext3 instead of ext4. Same result > every time. Switch to a different terminal and check the contents of the log files in /var/log. Let me know if you fine anything. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 30 December 2015 at 13:28, John Paul Adrian Glaubitz < glaub...@physik.fu-berlin.de> wrote: > On 12/30/2015 01:18 PM, Knut Petter Ølberg wrote: > > Which seems to work fine, until I get to the formatting part, where it > > seems to hang at 33%. This is in an LDOM on a T5140, tried changing from > > zvol to file, also tried ext2 and ext3 instead of ext4. Same result > > every time. > > Switch to a different terminal and check the contents of the log files > in /var/log. Let me know if you fine anything. > > Adrian > > -- > .''`. John Paul Adrian Glaubitz > : :' : Debian Developer - glaub...@debian.org > `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de > `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 > Seems to be hanging completely. I'll try on a T1000 later tonight. Maybe do the install on that one, and just transfer the image/zvol after installation. -kp
Re: New working sparc64 netinst image
On Wed, Dec 30, 2015 at 4:08 PM, Knut Petter Ølbergwrote: > On 30 December 2015 at 13:28, John Paul Adrian Glaubitz > wrote: >> >> On 12/30/2015 01:18 PM, Knut Petter Ølberg wrote: >> > Which seems to work fine, until I get to the formatting part, where it >> > seems to hang at 33%. This is in an LDOM on a T5140, tried changing from >> > zvol to file, also tried ext2 and ext3 instead of ext4. Same result >> > every time. >> >> Switch to a different terminal and check the contents of the log files >> in /var/log. Let me know if you fine anything. >> > Seems to be hanging completely. I'll try on a T1000 later tonight. Maybe do > the install on that one, and just transfer the image/zvol after > installation. well, i have had the same issue with 33% hanging formatting first partition. ┌─┤ Partitions formatting ├───┐ │ │ │ 33% │ │ │ │ Creating ext3 file system for /boot in partition #1 of /dev/vdiska... │ │ │ ─┘ and it's not hanging completely, i'm using network console in preseed (to have second console) d-i anna/choose_modules string network-console d-i network-console/password password r00tme d-i network-console/password-again password r00tme on the second console, in /var/log/syslog: Dec 30 14:32:51 partman: mke2fs 1.42.13 (17-May-2015) Dec 30 14:32:51 partman: Found a sun partition table in /dev/vdiska1 and in running processes: ~ # ps ... 10544 root 3464 S{50format_ext3} /bin/sh /lib/partman/commit.d/50form 10567 root 4008 Slog-output -t partman --pass-stdout mkfs.ext3 /dev/v 10568 root 9888 Smkfs.ext3 /dev/vdiska1 10602 root 4144 Sudpkg --configure --force-configure di-utils-shell 10603 root 3328 S{di-utils-shell.} /bin/sh /var/lib/dpkg/info/di-util 10604 root 3328 S{start-shell} /bin/sh /bin/start-shell di-utils-shel 10605 root 3320 Ssh -c /bin/sh 10606 root 3328 S/bin/sh 11396 root 3328 Rps My current oracle linux sparc and debian sparc64 installation, both have the following partition layout: (oracle sparc linux) : [root@linuxsparc mator]# parted /dev/vdiska GNU Parted 2.1 Using /dev/vdiska Welcome to GNU Parted! Type 'help' to view a list of commands. (parted) p Model: Unknown (unknown) Disk /dev/vdiska: 10.7GB Sector size (logical/physical): 512B/512B Partition Table: gpt Number Start End SizeFile system Name Flags 1 1049kB 1075MB 1074MB ext3 boot 2 1075MB 1076MB 1049kB bios_grub 3 1076MB 10.7GB 9661MB lvm (debian unstable sparc64 , installed from debian-9.0-sparc64-NETINST-1.iso ) : root@deb-t52:/home/mator# parted /dev/vdiska GNU Parted 3.2 Using /dev/vdiska Welcome to GNU Parted! Type 'help' to view a list of commands. (parted) p Model: Unknown (unknown) Disk /dev/vdiska: 10.7GB Sector size (logical/physical): 512B/512B Partition Table: gpt Disk Flags: Number Start End SizeFile system Name Flags 1 1049kB 1075MB 1074MB ext3 boot, esp 2 1075MB 1076MB 1049kBbios_grub 3 1076MB 3075MB 2000MB linux-swap(v1) 4 3075MB 10.7GB 7661MB ext4 So, basic difference is "gpt" on working sparc linux (both oracle and debian) and d-i is doing "sun" partition table and hanging on 33% while trying to format. As mentioned on https://oss.oracle.com/linux-sparc/docs/RELEASE-NOTES-1.0-en.html , quote "Disk Used for Installation Must Not Contain a Sun VTOC Partition" Too bad debian installer does not have usual linux partition tools, like fdisk and/or parted, to make manual changes to partitions created, and I've no idea how to check partition table partman or change it with preseed. I don't know solaris workaround for this problem (fmthard , prtvtoc , format , all via lofiadm , utilities does not allow to work with files like linux do). If you're installing in LDOM, you could probably use linux to partition disk file with parted (parted /path/to/file , mklabel gpt , mkpart ... ) and then use this file as disk in LDOM. And for bare-metal / physical hardware installations, i believe, we need to patch debian installer for creating gpt partition table instead of sun. Hope this helps. PS: btw the same problem with hanging format (as well disk detection) was on unstable jessie with installation , before sparc was dropped from debian, when i tried to install debian spring 2015.
Re: New working sparc64 netinst image
On Wed, Dec 30, 2015 at 12:19 PM, John Paul Adrian Glaubitzwrote: > On 12/30/2015 03:29 AM, Anatoly Pugachev wrote: >> On Tue, Dec 29, 2015 at 3:39 PM, John Paul Adrian Glaubitz >> wrote: >>> On 12/29/2015 01:31 PM, John Paul Adrian Glaubitz wrote: http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso >> >> installed from this cd/iso version into LDOM, using local mirror and >> preseed configuration > > Awesome, great to hear! Were there any issues you encountered? first, i had problems with formatting as well, but reused / copied disk from oracle sparc linux with gpt disk partition table (description of this problem in my previous email). after solving problem with formatting, d-i did not installed kernel, and i need to install it manually "apt-get install linux-image-sparc64-smp" , changing silo.conf and updating bootloader. then , with my first try I've installed with use of lvm , but was unable to properly boot into installed debian , dropping to initramfs prompt (is it busybox ?), since lvm initialization, from initrd, ./scripts/local-top/lvm2 , command lvchange_activate() { lvm lvchange -aay -y --sysinit --ignoreskippedcluster "$@" } gave me "bus error", and i need to re-install without lvm, installation went ok . where do i report lvm "bus error" currently ? Just checked with my installed sparc64 debian, adding lvm2 package, even simple commands give "bus error": root@deb-t52:/boot# dpkg -l lvm2 Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend |/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad) ||/ Name Version Architecture Description +++-==-==-==-= ii lvm2 2.02.138-1 sparc64Linux Logical Volume Manager root@deb-t52:~# lvm lvs Bus error >> now i have fully 64bit unstable/sid with 4.3 kernel and systemd working. > > Great. Let me know if you want me to set this up as a buildd. Since > this machine is extremely fast, we could run several buildd instances > without any noticable slow-down. Adrian, yeah, lets make this as buildd, about 200 vcpus , 200Gb ram and 300Gb hdd avail for this task. root@deb-t52:~# free -m totalusedfree shared buff/cache available Mem: 194432 228 193255 312 949 193626 Swap: 1906 01906 root@deb-t52:~# lscpu Architecture: sparc64 CPU op-mode(s):32-bit, 64-bit Byte Order:Big Endian CPU(s):192 On-line CPU(s) list: 0-191 Thread(s) per core:8 Core(s) per socket:12 Socket(s): 2 Flags: sun4v
Re: New working sparc64 netinst image
On 12/29/2015 03:15 PM, John Paul Adrian Glaubitz wrote: > On 12/29/2015 03:07 PM, Knut Petter Ølberg wrote: >> Started installing in an LDOM, seems to be working fine. Came as far as >> specifying the mirror, where I'm a bit lost as to what I should specify >> to make reprepo create a local repo for stretch. > > Normally you should be able to just use a regular Debian mirror, but > when you use ftp.debian-ports.org/debian, it's not recognized as > a mirror. I'm working on a quick mini-howto for reprepro with apt-cacher-ng and it seems I have to rebuild the netinst image to have debian-installer set the default suite to "unstable" (currently points to "stretch"). Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/29/2015 06:31 PM, John Paul Adrian Glaubitz wrote: > On 12/29/2015 06:22 PM, John Paul Adrian Glaubitz wrote: >> Hmm, let me test the new image first once it has been built. Might >> be that we don't need a mirror after all. Could be that d-i couldn't >> just use ftp.debian-ports.org because the suite was still set to >> "stretch" which ftp.debian-ports.org doesn't have. > > I have changed the suite to "sid" now: Still doesn't work really. But if anyone else wants to try, you can also configure debian-installer with the help of a preseed.cfg, see here for an example: > http://backup.parisc-linux.org/preseed-alpha.cfg Adopt for your own needs and arch. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/29/2015 05:57 PM, Kieron Gillespie wrote: > I have a server sitting in the cloud currently not doing anything > useful. It has a fairly good Internet connection. If I am pointed to > some instructions I could set up a dedicated mirror for SPARC64. Sounds good. Let me just test everything first. But in case you want to start already, here's the mini howto: $ apt-get install reprepro apt-cacher-ng $ mkdir /srv/debian-sparc64-archive $ cd /srv/debian-sparc64-archive $ wget https://people.debian.org/~glaubitz/reprepro-conf-sparc64.tgz $ tar xf reprepro-conf-sparc64.tgz $ reprepro -V update # this takes some time $ edit /etc/apt-cacher-ng/acng.conf Uncomment the following line: Port:3142 and under LocalDirs, add: LocalDirs: debian /srv/debian-sparc64-archive save the file and restart apt-cacher-ng: $ /etc/init.d/apt-cacher-ng restart # alternatively: systemctl restart apt-cacher-ng.service This should make the mirror available as "http://yourserver:3141/debian;. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/29/2015 06:05 PM, John Paul Adrian Glaubitz wrote: > On 12/29/2015 05:57 PM, Kieron Gillespie wrote: >> I have a server sitting in the cloud currently not doing anything >> useful. It has a fairly good Internet connection. If I am pointed to >> some instructions I could set up a dedicated mirror for SPARC64. > > Sounds good. Let me just test everything first. > > But in case you want to start already, here's the mini howto: Hmm, let me test the new image first once it has been built. Might be that we don't need a mirror after all. Could be that d-i couldn't just use ftp.debian-ports.org because the suite was still set to "stretch" which ftp.debian-ports.org doesn't have. Looking at /var/log/syslog when d-i runs, helps to debug this. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/29/2015 06:22 PM, John Paul Adrian Glaubitz wrote: > Hmm, let me test the new image first once it has been built. Might > be that we don't need a mirror after all. Could be that d-i couldn't > just use ftp.debian-ports.org because the suite was still set to > "stretch" which ftp.debian-ports.org doesn't have. I have changed the suite to "sid" now: > https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
I have a server sitting in the cloud currently not doing anything useful. It has a fairly good Internet connection. If I am pointed to some instructions I could set up a dedicated mirror for SPARC64. On Tue, Dec 29, 2015 at 11:53 AM, John Paul Adrian Glaubitz < glaub...@physik.fu-berlin.de> wrote: > On 12/29/2015 03:15 PM, John Paul Adrian Glaubitz wrote: > > On 12/29/2015 03:07 PM, Knut Petter Ølberg wrote: > >> Started installing in an LDOM, seems to be working fine. Came as far as > >> specifying the mirror, where I'm a bit lost as to what I should specify > >> to make reprepo create a local repo for stretch. > > > > Normally you should be able to just use a regular Debian mirror, but > > when you use ftp.debian-ports.org/debian, it's not recognized as > > a mirror. > > I'm working on a quick mini-howto for reprepro with apt-cacher-ng and > it seems I have to rebuild the netinst image to have debian-installer > set the default suite to "unstable" (currently points to "stretch"). > > Adrian > > -- > .''`. John Paul Adrian Glaubitz > : :' : Debian Developer - glaub...@debian.org > `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de > `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 > >
Re: New working sparc64 netinst image
Mark Morgan Lloyd wrote: John Paul Adrian Glaubitz wrote: On 12/29/2015 06:51 PM, John Paul Adrian Glaubitz wrote: Still doesn't work really. Ok, this seems to be a problem with the network configuration of my qemu setup. Can anybody report back who sits in front of real hardware? Damn- can anybody remind me how to get at the OBP settings from inside Solaris? I've got an Netra-20 reserved for this but it's auto-booting and I don't have a USB Sun keyboard with a Stop key. eeprom Booted off pukka CD-RW. This very earliest step was something that Debian has frequently got wrong, so I think it's worth testing. Several messages "EXT2 superblock magic is wrong" .. Debian installer main menu Language selection inoperative Keyboard selection works for British English Network devices detected Fails to allocate using DHCP (as expected) Am able to enter an IP4 address Correctly does a revers-lookup for the hostname Ditto for domain name Asks for manual entry of the Debian mirror Above all subject to the fact that I'm using a keyboard and screen rather than a serial terminal, so there could be details lost in transcription. I'm stopping there for the moment in the interest of supplying feedback promptly. Well done everybody :-) -- Mark Morgan Lloyd markMLl .AT. telemetry.co .DOT. uk [Opinions above are the author's, not those of his employers or colleagues]
Re: New working sparc64 netinst image
Hi, >As the debian infrastructure seems to be down at the moment, I have >uploaded my latest image to my webspace at work: > >> >http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso > >Feedback, please. Boots on a V240: ... Rebooting with command: boot cdrom Boot device: /pci@1e,60/ide@d/cdrom@0,0:f File and args: SILO Version 1.4.14 EXT2 superblock magic is wrong EXT2 superblock magic is wrong \ Welcome to Debian GNU/Linux sid! This is a Debian installation CDROM, built on 20151229-17:15. Keep it once you have installed your system, as you can boot from it to repair the system on your hard disk if that ever becomes necessary. WARNING: You should completely back up all of your hard disks before proceeding. The installation procedure can completely and irreversibly erase them! If you haven't made backups yet, remove the rescue CD from the drive and press L1-A to get back to the OpenBoot prompt. Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent permitted by applicable law. [ ENTER - Boot install ] [ Type "expert" - Boot into expert mode ] [ Type "rescue" - Boot into rescue mode ] boot: Allocated 64 Megs of memory at 0x4000 for kernel EXT2 superblock magic is wrong Loaded kernel version 4.3.3 EXT2 superblock magic is wrong ... And the kernel starts up, albeit a bit noisy with a lot of messages on the console: [ 127.261369] IRQ pre handler NOT supported. [ 127.261394] IRQ pre handler NOT supported. [ 127.261412] IRQ pre handler NOT supported. But the installer kicks in OK, so thumbs up! CU, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On Tue, Dec 29, 2015 at 3:39 PM, John Paul Adrian Glaubitzwrote: > On 12/29/2015 01:31 PM, John Paul Adrian Glaubitz wrote: >> http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso installed from this cd/iso version into LDOM, using local mirror and preseed configuration - spend too much time actually... now i have fully 64bit unstable/sid with 4.3 kernel and systemd working.
Re: New working sparc64 netinst image
On 29/12/15 19:43, John Paul Adrian Glaubitz wrote: > On 12/29/2015 06:51 PM, John Paul Adrian Glaubitz wrote: >> Still doesn't work really. > > Ok, this seems to be a problem with the network configuration of > my qemu setup. Can anybody report back who sits in front of real > hardware? Can you post your QEMU command line? By default you should get a slirp user-space stack that can talk to the outside world. ATB, Mark.
Re: New working sparc64 netinst image
John Paul Adrian Glaubitz wrote: On 12/29/2015 06:51 PM, John Paul Adrian Glaubitz wrote: Still doesn't work really. Ok, this seems to be a problem with the network configuration of my qemu setup. Can anybody report back who sits in front of real hardware? Damn- can anybody remind me how to get at the OBP settings from inside Solaris? I've got an Netra-20 reserved for this but it's auto-booting and I don't have a USB Sun keyboard with a Stop key. -- Mark Morgan Lloyd markMLl .AT. telemetry.co .DOT. uk [Opinions above are the author's, not those of his employers or colleagues]
Re: New working sparc64 netinst image
Hi, >I have changed the suite to "sid" now: > >> >https://people.debian.org/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso I think it's just the signature of the archive which the installer can't verify. From the log after the selection of debian-ports as mirror: Dec 29 21:04:35 main-menu[164]: INFO: Menu item 'choose-mirror' selected Dec 29 21:04:35 anna-install: Installing apt-mirror-setup Dec 29 21:04:35 anna[1221]: ERROR **: can't find packages file Dec 29 21:04:45 choose-mirror[1225]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/oldstable/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 29 21:04:46 choose-mirror[1225]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/stable/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 29 21:04:46 choose-mirror[1225]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/testing/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 29 21:04:46 choose-mirror[1225]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/unstable/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 29 21:04:47 choose-mirror[1225]: DEBUG: command: wget -q http://ftp.debian-ports.org/debian/dists/sid/Release -O - | grep -E '^(Suite|Codename|Architectures):' Dec 29 21:04:48 choose-mirror[1225]: DEBUG: command: wget -q http://ftp.debian-p--More-- (97% of 59444orts.org/debian//dists/sid/main/binary-sparc64/Release -O - | grep ^Architecture: Dec 29 21:04:48 anna-install: Installing sid-support Dec 29 21:04:48 anna[1247]: ERROR **: can't find packages file Dec 29 21:04:48 main-menu[164]: (process:1219): wget: server returned error: HTTP/1.1 404 Not Found Dec 29 21:04:48 main-menu[164]: (process:1219): wget: server returned error: HTTP/1.1 404 Not Found Dec 29 21:04:48 main-menu[164]: (process:1219): wget: server returned error: HTTP/1.1 404 Not Found Dec 29 21:04:48 main-menu[164]: DEBUG: resolver (libc6-udeb): package doesn't exist (ignored) Dec 29 21:04:48 main-menu[164]: INFO: Menu item 'download-installer' selected Dec 29 21:04:49 net-retriever: gpgv: Dec 29 21:04:49 net-retriever: Signature made Tue Dec 29 20:03:06 2015 UTC using RSA key ID C448326E Dec 29 21:04:49 net-retriever: gpgv: Dec 29 21:04:49 net-retriever: Can't check signature: public key not found Dec 29 21:04:49 net-retriever: error: Bad signature on /tmp/net-retriever-1275-Release. Dec 29 21:04:53 anna[1271]: WARNING **: bad d-i Packages file Dec 29 21:04:53 main-menu[164]: INFO: Menu item 'download-installer' succeeded but requested to be left unconfigured. Dec 29 21:04:53 main-menu[164]: DEBUG: resolver (libc6-udeb): package doesn't exist (ignored) Dec 29 21:04:56 main-menu[164]: INFO: Menu item 'di-utils-shell' selected Am I grateful for my serial console :-) But to me it looks like the installer is just missing a public key for debian-ports.org? CU, Uli -- Dipl. Inf. Ulrich Teichert|e-mail: ulrich.teich...@gmx.de | Listening to: Stormweg 24 |Imperial State Electric, The Estranged, 24539 Neumuenster, Germany|The About Blanks, Cyanide Pills, Fu Manchu.
Re: New working sparc64 netinst image
On 12/29/2015 08:53 PM, Mark Cave-Ayland wrote: > Can you post your QEMU command line? By default you should get a slirp > user-space stack that can talk to the outside world. $ qemu-system-sparc64 -net nic -net user -nographic -m 512 -cdrom ./debian-cd-created/debian-9.0-sparc64-NETINST-1.iso -boot d -hda debian.img The problem here is that the machine hosting this is in a university network and hosts don't randomly get IP addresses through DHCP, so I'd need some NAT setup or register the MAC of the qemu host. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 29/12/15 20:50, Mark Morgan Lloyd wrote: > Mark Morgan Lloyd wrote: >> John Paul Adrian Glaubitz wrote: >>> On 12/29/2015 06:51 PM, John Paul Adrian Glaubitz wrote: Still doesn't work really. >>> >>> Ok, this seems to be a problem with the network configuration of >>> my qemu setup. Can anybody report back who sits in front of real >>> hardware? >> >> Damn- can anybody remind me how to get at the OBP settings from inside >> Solaris? I've got an Netra-20 reserved for this but it's auto-booting and I >> don't have a USB Sun keyboard with a Stop key. > > eeprom > > Booted off pukka CD-RW. This very earliest step was something that Debian has > frequently got wrong, so I think it's worth testing. > > Several messages "EXT2 superblock magic is wrong" ^^ I wrote a reordering patch to hide that. It's entirely cosmetic since the error messages are a bit alarmist when they didn't need to be. silo-1.4.14-iso-before-ext2.patch diff -up silo-1.4.14/second/file.c.orig silo-1.4.14/second/file.c --- silo-1.4.14/second/file.c.orig 2013-03-19 16:52:22.859860434 -0400 +++ silo-1.4.14/second/file.c 2013-03-19 16:55:12.831856827 -0400 @@ -53,8 +53,8 @@ extern struct fs_ops ufs_fs_ops; /* Array of our supported ops */ static struct fs_ops *silo_fs_ops[] = { -_fs_ops, _fs_ops, +_fs_ops, _fs_ops, _fs_ops, NULL, >From solaris (and later from linux (eeprom is part of the prtconf package for linux)) eeprom auto-boot?=false setting this to false is usually enough to do the trick
Re: New working sparc64 netinst image
On 12/29/2015 10:48 PM, John Paul Adrian Glaubitz wrote: > On 12/29/2015 10:18 PM, Ulrich Teichert wrote: >> Am I grateful for my serial console :-) But to me it looks like the >> installer is just missing a public key for debian-ports.org? > > Which I don't really understand because I explicitly told debian > installer which keyring to use [1], namely debian-ports. Meh, I forgot. This keyring is just for checking the udebs... Anyway, you should be able to circumvent the issue with a preseed.cfg file where you can tell debian-installer to ignore the signature. Just clone and adapt the file for sparc64, put it up on some webspace and specify it on the kernel command line before booting the ISO. See: http://backup.parisc-linux.org/preseed-alpha.cfg Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/29/2015 06:51 PM, John Paul Adrian Glaubitz wrote: > Still doesn't work really. Ok, this seems to be a problem with the network configuration of my qemu setup. Can anybody report back who sits in front of real hardware? Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 12/29/2015 10:18 PM, Ulrich Teichert wrote: > Am I grateful for my serial console :-) But to me it looks like the > installer is just missing a public key for debian-ports.org? Which I don't really understand because I explicitly told debian installer which keyring to use [1], namely debian-ports. I'll look into this. Adrian > [1] http://anonscm.debian.org/cgit/d-i/debian-installer.git/commit/?id=32d62e3d9dd8b369801aee5aea7dec9b20644527 -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 29/12/15 20:01, John Paul Adrian Glaubitz wrote: > On 12/29/2015 08:53 PM, Mark Cave-Ayland wrote: >> Can you post your QEMU command line? By default you should get a slirp >> user-space stack that can talk to the outside world. > > $ qemu-system-sparc64 -net nic -net user -nographic -m 512 -cdrom > ./debian-cd-created/debian-9.0-sparc64-NETINST-1.iso -boot d -hda debian.img > > The problem here is that the machine hosting this is in a university > network and hosts don't randomly get IP addresses through DHCP, so > I'd need some NAT setup or register the MAC of the qemu host. The DHCP server is in-built - if you run dhclient or equivalent in the guest you should get a 10.0.2.15 IP address with a 10.0.2.2 gateway which is the host machine with the added ability to NAT outwards. Note that since this is userspace then unless you run QEMU as root then IIRC ping won't work, however normal TCP/UDP connections such as wget and telnet etc. should be fine. HTH, Mark.
New working sparc64 netinst image
Hi! As the debian infrastructure seems to be down at the moment, I have uploaded my latest image to my webspace at work: > http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso Feedback, please. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
Re: New working sparc64 netinst image
On 29/12/15 12:31, John Paul Adrian Glaubitz wrote: > Hi! > > As the debian infrastructure seems to be down at the moment, I have > uploaded my latest image to my webspace at work: > >> > http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso > > Feedback, please. A quick test with qemu-system-sparc64 (which is all I have to test with here) shows that the image now boots correctly all the way through to the Debian installer, although I need to increase the minimum amount of RAM to 256M compared with wheezy. Congratulations on reaching such an important milestone! Once this becomes an "official" Debian download, I'll add it to my OpenBIOS test suite to help shake out any other bugs with QEMU and the kernel. ATB, Mark.
Re: New working sparc64 netinst image
On 12/29/2015 03:07 PM, Knut Petter Ølberg wrote: > Started installing in an LDOM, seems to be working fine. Came as far as > specifying the mirror, where I'm a bit lost as to what I should specify > to make reprepo create a local repo for stretch. Normally you should be able to just use a regular Debian mirror, but when you use ftp.debian-ports.org/debian, it's not recognized as a mirror. Helge Deller (CC'ed) had the same issue on alpha and hppa and to resolve it, he created his own mirror using reprepro and made them available here: http://backup.parisc-linux.org/ (alpha and hppa only). Basically, creating a repo using reprepro and making it available using apt-cacher-ng should work. You can use the attached configuration for reprepro to create a valid sparc64 repository. Maybe someone create such a repository and make it available through the web via http. Just install reprepro, create a new folder and extract the contents of conf.tgz into it, then run "reprepro -V update" and wait. For apt-cacher-ng, this might help: > https://help.ubuntu.com/community/Apt-Cacher-Server Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 reprepro-conf-sparc64.tgz Description: application/compressed-tar
Re: New working sparc64 netinst image
On 29 December 2015 at 13:40, Mark Cave-Ayland < mark.cave-ayl...@ilande.co.uk> wrote: > On 29/12/15 12:31, John Paul Adrian Glaubitz wrote: > > > Hi! > > > > As the debian infrastructure seems to be down at the moment, I have > > uploaded my latest image to my webspace at work: > > > >> > > > http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso > > > > Feedback, please. > > A quick test with qemu-system-sparc64 (which is all I have to test with > here) shows that the image now boots correctly all the way through to > the Debian installer, although I need to increase the minimum amount of > RAM to 256M compared with wheezy. > > Congratulations on reaching such an important milestone! Once this > becomes an "official" Debian download, I'll add it to my OpenBIOS test > suite to help shake out any other bugs with QEMU and the kernel. > > > ATB, > > Mark. > > Started installing in an LDOM, seems to be working fine. Came as far as specifying the mirror, where I'm a bit lost as to what I should specify to make reprepo create a local repo for stretch. -kp
Re: New working sparc64 netinst image
On 12/29/2015 01:31 PM, John Paul Adrian Glaubitz wrote: > http://users.physik.fu-berlin.de/~glaubitz/debian-cd/debian-9.0-sparc64-NETINST-1.iso As a heads-up, you will probably create a local mirror from debian-ports which you then specify during installation: > https://parisc.wiki.kernel.org/index.php/How_to_create_Debian_unstable_iso_images#Create_a_local_copy_of_the_debian-ports_repository Specifying just ftp.debian-ports.org/debian apparently doesn't work. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913