=== denningsrogue1 is now known as denningsrogue
=== denningsrogue5 is now known as denningsrogue
=== zenlinux9 is now known as zenlinux
=== trekkie1701c_ is now known as trekkie1701c
[04:54] <cathode> hey all, ubuntu server 18.04 LTS running on a software md raid1 configuration, i did a dist upgrade to 20.04 and at the first reboot during the dist upgrade, the system now will not boot with a grub error... "Error: symbol 'grub_file_filters' not found."
[04:54] <cathode> and it dumps me to a grub recovery console.
[04:56] <tomreyn> !grub | cathode
[04:56] <ubot3> cathode: GRUB2 is the default Ubuntu boot manager. Lost GRUB after installing Windows? See https://help.ubuntu.com/community/RestoreGrub - For more information and troubleshooting for GRUB2 please refer to https://help.ubuntu.com/community/Grub2
[04:57] <cathode> the partitions on my pair of OS disks are ONLY the fd partitions
[04:57] <cathode> if i do grub-install against the disk will it destroy anything?
[04:58] <tomreyn> what are "fd partitions"?
[04:58] <cathode> linux raid autodetect type
[04:58] <tomreyn> but you have a partition table there? whjich type?
[04:59] <cathode> no, that's it
[04:59] <tomreyn> no partition table means you need to install grub elsewhere. maybe you had it embedded
[04:59] <tomreyn> but that was always sketchy and never recommended
[04:59] <cathode> nope. it was installed against the pair of OS disks only
[05:00] <tomreyn> bug 1891680
[05:00] <ubot3> Bug 1891680 in grub2 (Ubuntu Xenial) "grub-pc needs to detect when debconf points to invalid drive and stop in preinst, before unpacking files, and also treat this as a failure in postinst" [Critical, Confirmed] https://launchpad.net/bugs/1891680
[05:02] <tomreyn> i got there via dupe bug 1848797
[05:02] <ubot3> Bug 1891680 in grub2 (Ubuntu Xenial) "grub-pc needs to detect when debconf points to invalid drive and stop in preinst, before unpacking files, and also treat this as a failure in postinst" [Critical, Confirmed] [duplicate: 1848797] https://launchpad.net/bugs/1891680
[05:08] <cathode> I'm not really sure what to do with that
[05:28] <bipinsingh061> Hey anyone here ?
[05:30] <bipinsingh061> I have used ubuntu 20.04 before , it was a not so fast but ok , recently I got SSD installed on my laptop, and installed ubuntu 20.10 on laptop , its very slow , very laggy
[05:43] <tomreyn> bipinsingh061: laptop sounds like you'll be running a desktop there? join #ubuntu then
[05:43] <tomreyn> ah, you noticed, and did.
[07:34] <RoyK> is there a way to list kernels with kernel number, as in the number used by GRUB_DEFAULT? I'm not in front of this machine
[07:45] <TJ-> RoyK: what do you mean by 'kernel number'
[07:53] <RoyK> TJ-: from /etc/default/grub: GRUB_DEFAULT=0
[07:54] <RoyK> that can be changed, but I'd like a list of the valid ones before doing anything, avoiding to boot into nowhere
[07:56] <RoyK> since I don't have remote access to this box like ipmi/idrac or something and all work is from home because of covid19
[07:59] <TJ-> RoyK: that's a simple index number of the menuentry stanzas in /boot/grub/grub.cfg
[08:01] <TJ-> r
[08:01] <TJ-> RoyK: see https://www.gnu.org/software/grub/manual/grub/html_node/default.html#default
[10:58] <halvors> Is hwe available for the Raspberry Pi 4 image?
[11:11] <halvors> And how do i enable it?
[11:17] <juliank> halvors: no, rpi image uses linux-raspi kernel packages
[11:17] <juliank> waveform: ^
[11:24] <waveform> halvors, I'll try and find out from the kernel team what the long-term plan is with the focal kernel on the pi, but juliank is correct - at this time there's no hwe kernel package for the pi
[12:37] <waveform> halvors, the linux-raspi kernel is intended to be a "rolling" version (similar to the various cloud kernels, and hwe). The (tentative) plan at this stage is that on focal it'll skip 5.8 (groovy's kernel) and jump straight to hirsute's kernel version (once that is finalized)
[12:38] <halvors> waveform: Thank you so much :) And you too juliank. I'll wait for the hirsute.
[14:17] <whichpatrick> I'm trying to fix a VM but emergency.mode is booting with the disk in read-write mode and I cannot get it remounted in read-only.
[14:18] <whichpatrick> Is there some other way besides adding 'systemd.unit=emergency.mode' to the linux line in grub?
[14:18] <whichpatrick> Not sure what I'm doing wrong here, and googling for it just finds a bunch of results about how to get OUT of read-only.
[14:32] <teward> whichpatrick: have you tried `sudo mount -o remount,ro /` ?
[14:32] <teward> after you're booted.
[14:32] <teward> (force into ro mode)
[15:27] <whichpatrick> teward: yes, device is in use
[15:28] <whichpatrick> actually "mount point is busy"
[15:29] <whichpatrick> what's particularly confusing is that supposedly the difference between recovery mode and emergency mode is that emergency mode boots in ro
[15:29] <whichpatrick> but that's not what's happening :/
[15:49] <TJ-> whichpatrick: it's either just "emergency" or "systemd.unit=emergency.service"
[15:56] <teward> TJ-: how much do you know abotu systemd?
[15:56] <teward> and specifically mount type units
[15:57] <sdeziel> I thought it was "systemd.unit=emergency.target"
[16:00] <sdeziel> TIL: "emergency" is a shortcut
[16:14] <TJ-> sdeziel: there is a .target but strangely the docs talk about the .service being equiv to 'emergency'
[16:15] <TJ-> teward: how specific? I wrote a generator that created lots of bind mounts once
[16:16] <teward> TJ-: trying to create a template such unit that can do a specific local loop mount of a .img as part of another service - it's how i'm enforcing quotas for minio server instances.  (I have systemd units set up for the services, but have to /etc/fstab the mounts right now - would like to be able to create a template/mechanism that the instantiated units (minio@INSTANCE) can depend on so systemd knows it can/can't run the service without
[16:16] <teward> the correpsonding mount for that named instance)
[16:16] <teward> unfortunately... hitting the wall with systemd mount units.
[16:18] <sdeziel> TJ-: looks like the .service is activated by the .target
[16:19] <teward> creating the mount manually works fine
[16:19] <teward> creating the mount manually works fine (with `systemd-mount` and the proper arguments)
[16:19] <teward> trying to template it makes a whole mess of "can't find this object" or whatever in the systemd api/schema/objects
[16:19] <sdeziel> teward: you can see what was generated from /etc/fstab using: systemctl cat path-to-minio-img.mount
[16:21] <teward> sdeziel: i guess what i'm trying to do is shortcut having to create full-path mount names in the mount objects... and even if I can't, I can't make it a template to instantiate it.  I.E. minio-mounts-@.mount - which in THEORY can create an instantiated template mount unit - errors when referenced
[16:21] <teward> so I guess i'm having a headscratcher on how mount units differ
[16:22] <sdeziel> teward: what does your template look like?
[16:22] <teward> i'll pull it up in a sec, i need food *badly* or i'mma pass out.  *goes to fetch emergency food supplies*
[16:23] <ddstreet> teward from man systemd.mount "Note that mount units cannot be templated,"
[16:23] <teward> that explains that one
[16:23] <teward> so i'll just have to create individual mount units then...
[16:24] <ddstreet> yeah i think so
[16:25] <teward> ... wonder if i can cheat it with a `systemd-mount ...` call in ExecPreStart for the service... hmmmmmmmmm
[16:25] <teward> *goes to do testing, but first food*
[16:26] <TJ-> teward: or a specific generator (which can be just a shell script that writes them into /run/systemd/generator{,.late}/
[16:26] <TJ-> I can't find my generator now; must be on another system
[16:26] <teward> yeah i was gonna say I need examples xD
[16:30] <TJ-> teward: the *only* problem with this is, it doesn't work at boot-time but will any time later
[16:30] <TJ-> teward: due to, in my case, the LVMs it expects not being available at the time generators are executed
[16:30] <TJ-> teward: https://iam.tj/projects/ubuntu/systemd-generator-LVM_automount
[16:39] <TJ-> heading off; back later