FarO's questions

I would like to install Influxdb via a repo, not by installing manually each release, so I tried to add a repo for it, as I used to do with Influxdb v1 (I just updated my server after a long while).

I followed the instructions provided but "apt list influx*" gave no results:

# apt list influx*
Listing... Done

so I adapted it to ubuntu, by browsing the repo myself, so I after cleaning up things I repeated except by:

• changing from "-archive_compat" to "-archive", as recommended,

• skipping verification of the signature because I couldn't find the value to compare the "-archive" against

• change from debian to ubuntu

  wget -q https://repos.influxdata.com/influxdata-archive.key
  cat influxdata-archive.key | gpg --dearmor | sudo tee /etc/apt/trusted.gpg.d/influxdata-archive.gpg > /dev/null
  echo 'deb [signed-by=/etc/apt/trusted.gpg.d/influxdata-archive.gpg] https://repos.influxdata.com/ubuntu jammy stable' | sudo tee /etc/apt/sources.list.d/influxdata.list
  

and "apt update"

Now I get in /etc/apt/sources.list.d/influxdata.list:

deb [signed-by=/etc/apt/trusted.gpg.d/influxdata-archive.gpg] https://repos.influxdata.com/ubuntu jammy stable

however the same result:

# apt list influx*
Listing... Done

If I check manually https://repos.influxdata.com/ubuntu/dists/jammy/stable/binary-amd64/Packages I find there are packages inside it!

Where is my mistake?

I have an HP Elitebook 8460p from around 2012 which had Windows 10 on it. The BIOS is set to legacy, since UEFI is offered only for development purposes, with big warnings not to use it otherwise.

When I used a USB drive to "erase the disk and install" Ubuntu 21.10, everything went fine until the reboot, where I got "no operating system found" (error: 3F0). I checked with gparted and the disk was GPT with an UEFI partition.

This is strange, since Windows was likely in Legacy mode, given the BIOS settings, and also when I checked from the Live USB which mode is in use, I got legacy: there was no /sys/firmware/efi device.

I tried a lot of solutions found online, and at the end I was able to install 18.04 on the disk when I created a MBR partition table.

I didn't keep track of all tests I performed, but I am quite sure I also tried to install 21.10 AFTER putting a MBR partition table, so I'm confused.

So, how does the Ubuntu installer pick UEFI vs legacy when defining the way to install the system?

I would like to install 22.04 as soon as it is released and I would like to reformat using ZFS, but I need to know that it will work in Legacy mode, or I need to know how to make it work in Legacy mode.

There is another similar question (How to avoid UEFI? Installation without UEFI? How to have Ubuntu 19, 20, 21, 22 w/o UEFI? Just no more EFI/UEFI problems?) which is however about how to force Legacy. The solution proposed (MinimalCD) is outdated, from what I understand from the Ubuntu Wiki, deleting the EFI partition from the USB stick (I think) didn't work in my case, and also I'm mainly asking about the way Ubuntu picks the mode.

Edit

To be clear: in the BIOS the (U)EFI has never been active, the live media booted in Legacy, and yet the new install didn't start, showing the "no OS found" error. I tried to activate UEFI in the BIOS and the message did not show up anymore (just got a blank screen), leading me to think the install was in UEFI mode (also due to the UEFI partition).

I have a server 18.04 with custom ZFS kernel modules (0.8.x vs provided 0.7.x).

Periodically, when I update the kernel, I have to recompile the ZFS kernel module, which otherwise would get automatically disabled.

Last time I had to fiddle a while to get it done so this time I would like to know the correct steps involved, in advance.

I think

1. updating the kernel and kernel sources
2. downloading module source and compiling the module
3. removing previous kmod packages and installing the new ones

However, since the kernel module is required to access some filesystems which are currently mounted, what is the next step? simply reboot to have the new module mounted?

I also have, at some point, to enable the auto-import service, see https://unix.stackexchange.com/questions/338260/zfs-mounting-only-one-of-my-pools-on-boot

I have Ubuntu 18.04 and I installed minidlna via snap because of the several dependencies that I don't otherwise need, and to avoid dependency hell.

I edited /var/snap/minidlna-escoand/19/minidlna.conf and I checked in the log file that everything goes well:

[2019/05/01 18:42:06] minidlna.c:1048: warn: Starting MiniDLNA version 1.2.1.
[2019/05/01 18:42:06] minidlna.c:1089: warn: HTTP listening on port 8200

However, I checked with VLC and no media is available.

The config file shows:

media_dir=/tank/media/Videos
media_dir=/tank/media/Music
media_dir=/tank/media/Photos
media_dir=/tank/home/olaf/Anime e manga

and the permissions are:

root@ml110g7:~# ps aux | grep minidlna
root     ...  minidlnad -f /var/snap/minidlna-escoand/19/minidlna.conf -P /var/snap/minidlna-escoand/common/minidlna.pid
root@ml110g7:~# ls -l /tank/media/
total 27
drwxrws--- 13 root music  14 Feb 27 23:22 Music
drwxrws---  4 root photos  7 Feb 27 23:22 Photos
drwxrws---  9 root videos 10 Feb 27 23:22 Videos

minidlnad runs as root, so why isn't it able to access the folders? one of the folders contains spaces (something that should pose no problem), but the other ones are simpler and no error is logged.

How can I debug the issue?

I set up a network bonding for failsafe in Ubuntu 18.04 according to this netplan:

network:
    version: 2
    ethernets:
        eno1:
            dhcp4: no
        enp2s0:
            dhcp4: no
    bonds:
        bond0:
            addresses: [192.168.3.5/24]
            gateway4: 192.168.3.253
            nameservers:
                addresses: [208.67.222.222,208.67.220.220,8.8.8.8,8.8.4.4]
            dhcp4: no
            interfaces:
                - eno1
                - enp2s0
            parameters:
                mode: active-backup
                primary: eno1

Now when I log in I get:

IP address for enp2s0:  192.168.3.96
IP address for eno1:    192.168.3.96
IP address for bond0:   192.168.3.5

and ifconfig gives me (abridged):

bond0: flags=5187<UP,BROADCAST,RUNNING,MASTER,MULTICAST>  mtu 1500
        inet 192.168.3.5  netmask 255.255.255.0  broadcast 192.168.3.255
        inet6 xxx:a31d  prefixlen 64  scopeid 0x20<link>
        ether xxx:a3:1d  txqueuelen 1000  (Ethernet)

eno1: flags=6211<UP,BROADCAST,RUNNING,SLAVE,MULTICAST>  mtu 1500
        inet 192.168.3.96  netmask 255.255.255.0  broadcast 192.168.3.255
        ether xxx:a3:1d  txqueuelen 1000  (Ethernet)

enp2s0: flags=6211<UP,BROADCAST,RUNNING,SLAVE,MULTICAST>  mtu 1500
        inet 192.168.3.96  netmask 255.255.255.0  broadcast 192.168.3.255
        ether xxx:a3:1d  txqueuelen 1000  (Ethernet)

I am sure that, before the bond was set, one of the two MACs ended with :a3:1c

Did I make mistakes while setting the bond, or is this an intended behaviour?

Edit

I had to reinstall and this is the file that the installer generated for my same configuration, including the bond:

network:
    bonds:
        bond0:
            addresses:
            - 192.168.3.5/24
            gateway4: 192.168.3.253
            interfaces:
            - eno1
            - enp2s0
            nameservers:
                addresses:
                - 8.8.8.8
                - 8.8.4.4
            parameters:
                mode: active-backup
    ethernets:
        eno1: {}
        enp2s0: {}
    version: 2

I have a home server which needs a mail server for error messages, mostly.

I used postfix but I find the features/complexity compromises not optimal for my home server: there are tens of lines to check/configure, stunnel to install for TLS, ...

What simpler mail server with TLS support can I use, that is also packaged by Canonical? I'm going to only use an external relay server provided by my provider.

Since the first installation of my system I modified various config files in /etc and I customised my system.

Copying the whole /etc for storage in a safe location before reinstalling/upgrading would slow down and potentially damage a new install, therefore how can I collect all and only the modified files?