[DNG] Beowulf installation skipped networking. Quickest fix?
My shiny new Beowulf 3.0.0 install on an Intel NUC skipped ethernet stuff, perhaps because the ethernet cable wasn't inserted at the time? Now, I can start hacking /etc/network/interfaces, /etc/hosts, resolv.conf, etc., but will that reveal so much missing that I'm better off doing another install with the LAN plugged in? Heck, are there kernel modules missing? Erik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble booting Beowulf on an Intel NUC8i5BEK : SOLVED
Being stuck on a horrible webmail thing until the new host is up, I'll limit the number of posts lacking In-Reply-To for proper MUA threading, by replying to Marjorie, Steve, and Florian here, shortest first. TL;DR: Punchline at the end. On 25.03.21 14:28, Florian Zieboll via Dng wrote: > On Thu, 25 Mar 2021 17:30:14 +1030 dva...@internode.on.net wrote: > > > Have just taken delivery of an Intel NUC8i5BEK, installed RAM & SSD. > > A USB stick with: > > $ dd if=~/Downloads/devuan_beowulf_3.0.0_i386_desktop-live.iso > > of=/dev/sdb bs=512k > > looks good mounted on another host, but doesn't boot in the NUC. > > Does the "other host" recognize it as bootable in UEFI mode? Did you > check for a BIOS update? Did you try with "Secure Boot" disabled? Ah, have now tried with "Secure Boot" disabled. No change. And it doesn't boot the laptop, either. I'll give a unetbootin version a go. (As dd had worked with the Udoo X86, I'd become complacent.) See below. On 25.03.21 07:21, Steve Litt wrote: > My experience, over many years with many distros, is that "bootable > thumb drives" aren't so bootable. Some work with a dd transfer but > don't work with unetbootin. Others work with unetbootin but not dd > Others require all sorts of special sauce to get them bootable. You > mentioned that your thumb drive "looks good" on another host. Unless > "looks good" means the other host booted off it, the thumb drive's > bootability is still up in the air, and when combined with uefi/legacy, > boot order, and all sorts of other stuff, the last thing you need is a > boot media you can't trust. Yeah, the first time around that special sauce gets up my nose - probably from not holding my mouth right. Rats, the unetbootin version of Beowulf also didn't boot in either the NUC or the laptop. > I'd purchase a $35.00 USB DVD burner > drive, burn a bootable CD or DVD, then boot off of that I'm sorely tempted, though it isn't here now, so I'll persist a little more, as that's usually what it takes. (Achieved occupancy permit for a rural off-grid owner-build on Monday, after 2.5 years of effort, including self-drawn plans getting planning and building approval, and digging 7 m³ of trench by hand during a Covid lockdown, so am in a persistent mood.) On 25.03.21 10:57, Marjorie Roome via Dng wrote: > Hi Erik, > > I have an older, more basic NUC running in legacy mode as my email > server, however I think you have the newer AptioV bios. > > Check here, where you can see the two possible BIOS start pages: > > https://www.intel.com/content/www/us/en/support/articles/06028/intel-nuc.html It turns out that my NUC also has the Visual BIOS. It was pretty easy to navigate after the first time through, with more than one path to the "Advanced" boot stuff. [snip] > I think you need the investigate the 'Boot' tab first. Yes, and I'd love to try legacy mode, but that just tells me "No Boot Drive" when I select it in the Boot Order window's legacy tab. Oh Farnarkle! Do you know what the secret sauce is? Ignore the "No Boot Drive" report when enabling Legacy Boot. Disable UEFI Boot. Persist in pressing F10 to save the config, despite the BIOS ignoring the first three presses, and hanging for ten seconds or so. Mind you, my euphoria was somewhat dimmed by the fact that the long scroll of linux boot messages ended in a uniformly light grey screen devoid of any content, textual or graphical. However, ditching the unetbootin stick and returning to the dd version, all is good. It boots all the way to the Devuan desktop. Now I just have to get the battery corrosion off the neglected cordless mouse, so I can do clicky stuff. (Taking time out for long projects does have consequences.) Many thanks folks, for helping me find the special sauce on this one. A BIOS like that does tend to keep Linux within a motivated community, I figure. Erik (Rather pleased with progress so far. :) ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] Chimaera upgrade
The upgrade of my still libsystemd-free desktop PC from Ascii to Chimaera ~2 weeks ago went so smooth that I forgot to mention it: $ apt update && apt dist-upgrade $ sed -i\.ascii 's/ascii/chimaera/' /etc/apt/sources.list $ apt update && apt upgrade (to install: 223 / upgrade: 903) $ apt dist-upgrade (to install: 655 / upgrade: 628 / remove: 127) On the way, 'recollgui' threw an error (I didn't even remember, just rediscovered it in the logs when checking the package counts): Unpacking recollgui (1.28.5-1) over (1.24.3-3) ... [1mdpkg:[0m error processing archive /tmp/apt-dpkg-install-ks8aVf/20-recollgui_1.28.5-1_amd64.deb (--unpack): trying to overwrite '/usr/share/recoll/doc/docbook-xsl.css', which is also in package recollcmd 1.24.3-3 But it was easy to recover: $ apt --fix-broken install (to install: 40 / upgrade: 49 / remove: 34) $ apt dist-upgrade (to install: 235 / upgrade: 325 / remove: 65) $ reboot $ apt autoremove (to remove: 140) Thumbs up and thank you for this great distro! Florian ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] kvm -- virsh failed shutdown -- reason was missing acpid on vm [resolved]
On Thursday 25 March 2021 at 18:53:20, Andrew McGlashan via Dng wrote: > Hi, > > On 26/3/21 12:25 am, Antony Stone wrote: > > > > Indeed - this is a known requirement for VMs (certainly under KVM, I > > don't know about Xen etc, but I would assume it also applies there). > > I think that systemd takes care of it, with or without acpid, but I'm not > sure about that. I have no idea about that at all. > If this is a consequence of choosing non-systemd, then perhaps it should be > mitigated by the sans systemd system installing acpid. Sounds reasonable, but I wonder whether Debian Jessie-- automatically installed it? > > Whether you consider it a shortcoming of Debian (and therefore Devuan) > > that acpid and acpid-support-base are not installed by default is up to > > you. > > Debian, no need for acpid for vm but only if systemd is installed, is that > correct? I have no familiarity with machines running systemd. > > My solution is that I've added these packages to the script I run > > immediately after creating a VM, to make sure it works the way I want it > > to. > > Yeah, it was a pain point until I worked it out. I agree. Antony. -- "It wouldn't be a good idea to talk about him behind his back in front of him." - murble Please reply to the list; please *don't* CC me. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] kvm -- virsh failed shutdown -- reason was missing acpid on vm [resolved]
Hi, On 26/3/21 12:25 am, Antony Stone wrote: > On Thursday 25 March 2021 at 14:16:52, Andrew McGlashan via Dng wrote: > >> "virsh shutdown vmname" >> >> The immediate response was that it would shutdown the vm >> >> However, doing "virsh list --all" still showed the vm as "Running" no >> matter how long I waited. >> >> It turns out the the vm needed to have the acpid package installed so that >> the vm could get the shutdown signal. > > Indeed - this is a known requirement for VMs (certainly under KVM, I don't > know about Xen etc, but I would assume it also applies there). I think that systemd takes care of it, with or without acpid, but I'm not sure about that. If this is a consequence of choosing non-systemd, then perhaps it should be mitigated by the sans systemd system installing acpid. > Whether you consider it a shortcoming of Debian (and therefore Devuan) that > acpid and acpid-support-base are not installed by default is up to you. Debian, no need for acpid for vm but only if systemd is installed, is that correct? > My solution is that I've added these packages to the script I run immediately > after creating a VM, to make sure it works the way I want it to. Yeah, it was a pain point until I worked it out. Cheers A. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble booting Beowulf on an Intel NUC8i5BEK
On Thu, 25 Mar 2021 17:30:14 +1030 dva...@internode.on.net wrote: > Have just taken delivery of an Intel NUC8i5BEK, installed RAM & SSD. > A USB stick with: > $ dd if=~/Downloads/devuan_beowulf_3.0.0_i386_desktop-live.iso > of=/dev/sdb bs=512k > looks good mounted on another host, but doesn't boot in the NUC. Does the "other host" recognize it as bootable in UEFI mode? Did you check for a BIOS update? Did you try with "Secure Boot" disabled? libre Grüße, Florian ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] kvm -- virsh failed shutdown -- reason was missing acpid on vm [resolved]
On Thursday 25 March 2021 at 14:16:52, Andrew McGlashan via Dng wrote: > "virsh shutdown vmname" > > The immediate response was that it would shutdown the vm > > However, doing "virsh list --all" still showed the vm as "Running" no > matter how long I waited. > > It turns out the the vm needed to have the acpid package installed so that > the vm could get the shutdown signal. Indeed - this is a known requirement for VMs (certainly under KVM, I don't know about Xen etc, but I would assume it also applies there). https://serverfault.com/questions/441204/kvm-qemu-guest-shutdown-problems Whether you consider it a shortcoming of Debian (and therefore Devuan) that acpid and acpid-support-base are not installed by default is up to you. My solution is that I've added these packages to the script I run immediately after creating a VM, to make sure it works the way I want it to. Antony. -- I conclude that there are two ways of constructing a software design: One way is to make it so simple that there are _obviously_ no deficiencies, and the other way is to make it so complicated that there are no _obvious_ deficiencies. - C A R Hoare Please reply to the list; please *don't* CC me. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] kvm -- virsh failed shutdown -- reason was missing acpid on vm [resolved]
Hi, I set up a new kvm machine with beowulf 3.1 and created a simple vm that I plan to use for Wireguard. The problem I had was with doing: "virsh shutdown vmname" The immediate response was that it would shutdown the vm However, doing "virsh list --all" still showed the vm as "Running" no matter how long I waited. It turns out the the vm needed to have the acpid package installed so that the vm could get the shutdown signal. Kind Regards AndrewM -- Andrew McGlashan ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble booting Beowulf on an Intel NUC8i5BEK
dva...@internode.on.net said on Thu, 25 Mar 2021 17:30:14 +1030 > Have just taken delivery of an Intel NUC8i5BEK, installed RAM & SSD. >A USB stick with: >$ dd if=~/Downloads/devuan_beowulf_3.0.0_i386_desktop-live.iso >of=/dev/sdb bs=512k >looks good mounted on another host, but doesn't boot in the NUC. > >On power-up or Ctl-Alt-Del, the NUC reads the USB twice (LED on the >stick flashes), >but then reports "A bootable device has not been detected." My experience, over many years with many distros, is that "bootable thumb drives" aren't so bootable. Some work with a dd transfer but don't work with unetbootin. Others work with unetbootin but not dd. Others require all sorts of special sauce to get them bootable. You mentioned that your thumb drive "looks good" on another host. Unless "looks good" means the other host booted off it, the thumb drive's bootability is still up in the air, and when combined with uefi/legacy, boot order, and all sorts of other stuff, the last thing you need is a boot media you can't trust. I'd purchase a $35.00 USB DVD burner drive, burn a bootable CD or DVD, then boot off of that. SteveT Steve Litt Spring 2021 featured book: Troubleshooting Techniques of the Successful Technologist http://www.troubleshooters.com/techniques ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Trouble booting Beowulf on an Intel NUC8i5BEK
Hi Erik, On Thu, 2021-03-25 at 17:30 +1030, dva...@internode.on.net wrote: > > Have just taken delivery of an Intel NUC8i5BEK, installed RAM & SSD. > A USB stick with: > $ dd if=~/Downloads/devuan_beowulf_3.0.0_i386_desktop-live.iso > of=/dev/sdb bs=512k looks good mounted on another host, but doesn't > boot in the NUC. > > On power-up or Ctl-Alt-Del, the NUC reads the USB twice (LED on the > stick flashes), but then reports "A bootable device has not been > detected." > > I've gone into the BIOS, and in the Boot Order tab, hit Advanced, > then in Boot Configuration tab set "Boot USB Devices First". F10 to > save. > That gives a Boot Devices list below with a greyed out "Internal UEFI > Shell", > followed by "USB", then the others below that. Looks good, but > returning to the adjacent Boot Priority tab shows: > > UEFI Boot Priority Legacy Boot Priority > > --- > UEFI Boot Legacy Boot > Boot Drive Order Boot Drive Order "No > Boot Drive" > UEFI LAN: PXE IP4 ... > UEFI LAN: PXE IP6 ... > > and the Boot Order tab seen on entry to the BIOS shows the same. > > From that I infer that legacy boot from USB is not a BIOS config > option, and that the Boot Configuration tab setting isn't recognised > in the Boot Priority tab's UEFI column. (At least for display, since > the USB is accessed twice during boot.) > > Can anyone spread a sprinkle of enlightenment on this apparent > impasse? > > Erik > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng I have an older, more basic NUC running in legacy mode as my email server, however I think you have the newer AptioV bios. Check here, where you can see the two possible BIOS start pages: https://www.intel.com/content/www/us/en/support/articles/06028/intel-nuc.html Looking in the AptioV Glossary: https://www.intel.com/content/dam/support/us/en/documents/intel-nuc/NUC-AptioV-BIOS-Glossary.pdf I think you need the investigate the 'Boot' tab first. -- Marjorie ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] Trouble booting Beowulf on an Intel NUC8i5BEK
Have just taken delivery of an Intel NUC8i5BEK, installed RAM & SSD. A USB stick with: $ dd if=~/Downloads/devuan_beowulf_3.0.0_i386_desktop-live.iso of=/dev/sdb bs=512k looks good mounted on another host, but doesn't boot in the NUC. On power-up or Ctl-Alt-Del, the NUC reads the USB twice (LED on the stick flashes), but then reports "A bootable device has not been detected." I've gone into the BIOS, and in the Boot Order tab, hit Advanced, then in Boot Configuration tab set "Boot USB Devices First". F10 to save. That gives a Boot Devices list below with a greyed out "Internal UEFI Shell", followed by "USB", then the others below that. Looks good, but returning to the adjacent Boot Priority tab shows: UEFI Boot Priority Legacy Boot Priority --- UEFI Boot Legacy Boot Boot Drive Order Boot Drive Order "No Boot Drive" UEFI LAN: PXE IP4 ... UEFI LAN: PXE IP6 ... and the Boot Order tab seen on entry to the BIOS shows the same. From that I infer that legacy boot from USB is not a BIOS config option, and that the Boot Configuration tab setting isn't recognised in the Boot Priority tab's UEFI column. (At least for display, since the USB is accessed twice during boot.) Can anyone spread a sprinkle of enlightenment on this apparent impasse? Erik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
