I am using Singularity on RHEL-8. I just upgraded to cgroupsv2 so that I could throttle resource usage. I also delegate CPU controller permissions in my /etc/systemd/system/user@.service.d/override.conf as described [here](https://github.com/systemd/systemd/issues/18293) . So far, so good. So I try to run: singularity run --cpus 0.8 container.sif. I get the following error (some of the random strings randomized and shortened): > FATAL: container creation failed: while applying cgroups config: while creating cgroup manager: failed to connect to dbus (hint: for rootless containers, maybe you need to install dbus-user-session package, see https://github.com/opencontainers/runc/blob/master/docs/cgroup-v2.md) : could not execute busctl --user --no-pager status (output: "Failed to get credentials: No such process\nBusAddress=unix:abstract=/tmp/dbus-4843,guid=35345\nBusScope=user\nBusID=343345\n"): exit status 1 Indeed, if I try busctl --user status, I get a connection refused error. There are some related questions like [this one](https://superuser.com/questions/1561076/systemctl-user-failed-to-connect-to-bus-no-such-file-or-directory-debian-9) and [this one](https://unix.stackexchange.com/questions/615917/failed-to-get-d-bus-connection-connection-refused) , but my dbus session seems to be running fine, and I am not trying to run this as root. What gives?

I have a BASH pipeline which at a point runs a Singularity container with *singularity exec* as follows:

singularity exec --bind pwd:/folder --bind $d:/results .sif  -i /folder/.fastq -v /results//.vcf -r /folder/.fna -s  -j 24 -t 24 -o /results/

Since I'm running multiple experiments at once with an array, I'm redefining the experiments with an environment variable that I wish to add to the `; it works for all the steps of the pipeline but Singularity` seems it's unable to *see* the variable(s) I'm defining in my script... Does anyone have any advice on how to do so? I looked up a bit but the --env does not seem to be working. Thanks in advance!

I am building a Singularity container using Apptainer. I installed package libfakeroot using command apt install -y libfakeroot in the definition file in hope that /.singularity.d/libs/libfakeroot.so could be seen by normal user. But because the user is (fake) root when the package libfakeroot is installed, this file is only visible when I shell into the container as a fake root user (shell --fakeroot). If I shell into the container as a normal user (without --fakeroot), the directory /.singularity.d/libs/ is empty. Is there any way to make /.singularity.d/libs/libfakeroot.so visible to fakeroot also visible to a normal user? Thank you. Environment: - Host OS: RedHat 8.8 - Guest OS: Ubuntu 22.04 - Apptainer version: 1.2.5

We have several machines that we want to run the same Singularity container on. As we already have a NAS mounted by every machine, we thought this would be a good place to store it without the need to think about distribution. Initially, we put a read-only container to test it, which was successful. We then needed to install additional system packages inside the container. Thus, we build a new sandbox container directly on the NAS (we have root permission on the NAS). However, we were not able to install packages in the sandbox (the singularity shell was started with the write flag

-w

and root permissions) and got the following error. E.g., from starting

update

inside the sandbox: error: cannot get shared lock on /var/lib/rpm/Packages error: cannot open Packages index using db5 - Operation not permitted (1) error: cannot open Packages database in /var/lib/rpm Error: Error: rpmdb open failed When we build the sandbox container locally on one of the machines and access it, we can install software through the package manager. To setup we do these commands. The first builds the sandbox from the library of images. In this case AlmaLinux 8 (we also tried it with CentOS 7, but it has the same problem). Then we enter the sandbox in shell mode to install packages. singularity build --sandbox al8.6-sandbox library://library/default/almalinux:8 singularity shell -w -B /mnt:/mnt al8.6-sandbox What can be the cause for the error on the NAS?

In Singularity when building under fake-root anything Ubuntu based with in %post either apt update -y or apt install -y openssh-client or their apt-get versions or with -f in the definition file fails fatally with groupadd: failure while writing changes to /etc/group addgroup: `/sbin/groupadd -g 101 ssh' returned error code 10. Exiting. Errors were encountered while processing: openssh-client E: Sub-process /usr/bin/dpkg returned an error code (1) FATAL: While performing build: while running engine: exit status 100 I see why adding a user is an issue, but yum update works fine in Rocky 9 (I cannot switch to it due to some requirements). This is on a cluster where I do NOT have root priviledges and have to use Singularity and docker isn't available —the base OS is CentOS 7 and glibc issues are insurmountable. I have tried a bunch of different Ubuntu images, ranging from Nvidia's Docker Ubuntu image and officially signed to unsigned Singularity images. Conda can install ssh, but the issue is apt update fails and I cannot work around it. I feel like I missing something obvious based on the lack of much on this online...

I constructed a singularity/apptainer sif container and I can now start an instance of this container. Is there a way to grant other users (in a group) the right to run the instance ? In short, How can I makes things such that any user in the group users is able to do singularity run instance:\\myinstance