Marking as "invalid" as it's not a bug in nvidia-graphics-drivers-510
but an external NVIDIA component.
** Changed in: nvidia-graphics-drivers-510 (Ubuntu)
Status: Confirmed => Invalid
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to nvidia-graphics-drivers-510 in Ubuntu.
https://bugs.launchpad.net/bugs/1971978
Title:
Driver binaries fail to load on arm64 through LXD
Status in nvidia-graphics-drivers-510 package in Ubuntu:
Invalid
Bug description:
On a fresh arm64 system I failed to start LXD containers with NVIDIA
GPUs attached. The error message LXD / nvidia-container-cli fails with
is:
lxc t0 20220506112123.475 DEBUG conf - conf.c:run_buffer:310 - Script exec
/snap/lxd/current/lxc/hooks/nvidia produced output: + exec nvidia-container-cli
--user configure --no-cgroups --ldconfig=@/usr/sbin/ldconfig.real --compute
--utility --require= --require= /var/snap/lxd/common/lxc/
lxc t0 20220506112123.492 DEBUG conf - conf.c:run_buffer:310 - Script exec
/snap/lxd/current/lxc/hooks/nvidia produced output: nvidia-container-cli.real:
initialization error: load library failed: /lib/aarch64-linux-gnu/libc.so.6:
version `glibc_2.35' not found (required by
/lib/aarch64-linux-gnu/libgcc_s.so.1)
This seems to be related to how the aarch64 driver is linked and
cannot be reproduced on amd64 on Ubuntu 22.04 with the same packages
installed as on arm64.
The error message indicates it's not nvidia-container-cli failing to
start but rather a dependency it fails to load due to missing symbols
in the core20 libc the LXD snap uses.
Steps to reproduce:
1. Setup a fresh Ubuntu 22.04 arm64 system with an NVIDIA GPU attached
2. Install the 510 driver from the archive
$ sudo apt install nvidia-utils-510 linux-modules-nvidia-510-generic
libnvidia-compute-510 --purge
3. Create a container and configure it
$ lxc init ubuntu:f c0
# Adjust the PCI address to match your GPU
$ lxc config device add c0 gpu gpu pci=0000:01:00.0
$ lxc config set c0 nvidia.runtime=true
# Make logs a bit more verbose
$ lxc config set c0 raw.lxc lxc.log.level=0
4. Now start the container and see it failing
$ lxc start c0
Error: Failed to run: /snap/lxd/current/bin/lxd forkstart c0
/var/snap/lxd/common/lxd/containers /var/snap/lxd/common/lxd/logs/c0/lxc.conf:
Try `lxc info --show-log c0` for more info
5. Check logs and see the error mentioned above:
$ lxc info --show-log c0
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-510/+bug/1971978/+subscriptions
--
Mailing list: https://launchpad.net/~kernel-packages
Post to : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help : https://help.launchpad.net/ListHelp
