Closed btreut closed 1 year ago
Does it make sense to re-create the box with --init
?
Does it make sense to re-create the Debian box with the --init
option?
Hi @bsvh You can try doing an init distrobox But be aware, only opensuse and the other distros with "-init" in their image name will work, no Debian with init support is there
@89luca89 thanks for
"-init" in their image name will work, no Debian with init support is there
before switching to a completely other distro, I started investigating the situation a bit and found that entering the Debian box and issuing a pstree -p
reveals a systemd process with pid 1 exactly the same as the one found on the host (pstree-distrobox-debian.txt, pstree-eos.txt). So I am puzzled: why does RStudio complain and why is it unable to communicate with pid 1? More info about my system (including process list under heading .= Processes =
can be found in the eos-diagnostic-230326_163444_UTC 0200.txt file linked in my post on the EndlessOS community platform.
ok, maybe the systemd init in the box is a dead end. I started digging a bit more and found the following:
in the box: there is no /run/dbus/system_bus_socket
, but this socket is available in the host environment.
What is the best way to allow access in the box to the socket /run/dbus/system_bus_socket
?
So you have systemd running on your host?
What command did you use for creating the box?
I personally never got these problems, but I am on Fedora Kinoite with systemd and use various distroboxes created just as is.
So you have systemd running on your host?
absolutely correct
What command did you use for creating the box?
distrobox create --image debian:latest --name debian
and @89luca89 does not recommend to add a -I
if the image is not prepared for that.
What is the best way to allow access in the box to the socket
/run/dbus/system_bus_socket
?
Found this post on stackoverflow, does this provide a solution?
And: how can I specify the -v/--volume
option for the podman command in the distrobox-enter command?
Very strange: I followed the instructions of @firefoxlover given in this post and there are no problems with communication with systemd
under Fedora.
So this might be an issue with the Debian image, therefore I close this issue here.
.
yes the images sometimes have weird problems, I once had Ubuntu without apt
I also would use the newest image possible e.g. not Debian. Fedora with the COPR for the modules should work, but I dont know how other distributions handle that
I am trying to get RStudio running in the OStree/flatpak based distribution EndlessOS, finnaly arrived at a solution based on podman (see post). With a newer version of EndlessOS, podman was upgraded and I restarted experimenting this time with distrobox. I successfully set up a Debian bullseye (11.6) container and installed RStudio within the container ignoring the following messages I got during startup of Rstudio:
since RStudion did run without apparent problems. But after installation of several RStudio extensions (and some deb packages, which were required for installation of these extensions) for learning R (shiny,curl,libcurl,learnr) I encountered during installation of tidyverse the following error within RStudio:
Reading a bit in Podman in Action about systemd integration, I fear that setting this up correctly is beyond my knowledge, so I am asking here for help.
Additional Information: podman version 3.0.1 (it is afaik tightly integrated in Endless OS 5.0.1) distrobox version 1.4.2.1