Collabora-code app container not starting since updating incus to 7.3

pachulo · August 4, 2026, 10:39pm

Since the latest Incus update (7.3) I can’t get my collabora-code app container started.
This is what I see on the logs:

$ incus info --show-log collabora-code-container
Name: collabora-code-container
Description:
Status: STOPPED
Type: container (application)
Architecture: aarch64
Created: 2025/08/29 23:55 UTC
Last Used: 2026/08/04 07:46 UTC

Log (lxc.log):

lxc collabora-code-container 20260804074625.331 ERROR    utils - ../src/lxc/utils.c:open_without_symlink:1196 - Too many levels of symbolic links - hosts in /opt/incus/lib/lxc/rootfs/etc/hosts was a symbolic link!
lxc collabora-code-container 20260804074625.331 ERROR    conf - ../src/lxc/conf.c:mount_entry:2101 - Too many levels of symbolic links - Failed to mount "/var/lib/incus/containers/collabora-code-container/network/hosts" on "/opt/incus/lib/lxc/rootfs/etc/hosts"
lxc collabora-code-container 20260804074625.331 ERROR    conf - ../src/lxc/conf.c:lxc_setup:3820 - Failed to setup mount entries
lxc collabora-code-container 20260804074625.331 ERROR    start - ../src/lxc/start.c:do_start:1466 - Failed to setup container "collabora-code-container"
lxc collabora-code-container 20260804074625.332 ERROR    sync - ../src/lxc/sync.c:sync_wait:34 - An error occurred in another process (expected sequence number 3)
lxc collabora-code-container 20260804074625.337 ERROR    network - ../src/lxc/network.c:lxc_netdev_restore_altnames:1422 - Invalid argument - Failed to get altnames for interface "vethd837a0d0"
lxc collabora-code-container 20260804074625.337 WARN     network - ../src/lxc/network.c:lxc_delete_network_priv:3940 - Failed to restore altnames for interface with index 0 and initial name "vethd837a0d0"
lxc collabora-code-container 20260804074625.337 WARN     network - ../src/lxc/network.c:lxc_delete_network_priv:3945 - Failed to rename interface with index 0 from "physdrGajQ" to its initial name "vethd837a0d0"
lxc collabora-code-container 20260804074625.338 ERROR    lxccontainer - ../src/lxc/lxccontainer.c:wait_on_daemonized_start:837 - Received container state "ABORTING" instead of "RUNNING"
lxc collabora-code-container 20260804074625.338 ERROR    start - ../src/lxc/start.c:__lxc_start:2288 - Failed to spawn container "collabora-code-container"
lxc collabora-code-container 20260804074625.338 WARN     start - ../src/lxc/start.c:lxc_abort:1229 - No such process - Failed to send SIGKILL via pidfd 17 for process 3176
lxc 20260804074625.438 ERROR    af_unix - ../src/lxc/af_unix.c:lxc_abstract_unix_recv_fds_iov:218 - Connection reset by peer - Failed to receive response
lxc 20260804074625.438 ERROR    af_unix - ../src/lxc/af_unix.c:lxc_abstract_unix_recv_fds_iov:218 - Connection reset by peer - Failed to receive response
lxc 20260804074625.438 ERROR    commands - ../src/lxc/commands.c:lxc_cmd_rsp_recv_fds:128 - Failed to receive file descriptors for command "get_state"
lxc 20260804074625.438 ERROR    commands - ../src/lxc/commands.c:lxc_cmd_rsp_recv_fds:128 - Failed to receive file descriptors for command "get_init_pid"

The container was created this way:

incus launch docker:collabora/code:latest collabora-code-container

What can I do to solve the issue and start it?

Thanks!

stgraber · August 5, 2026, 4:31am

incus file remove collabora-code-container/etc/hosts
incus start collabora-code-container

stgraber · August 5, 2026, 4:31am

It looks like their /etc/hosts is a symlink, which then trips some of our security checks in liblxc.

pachulo · August 5, 2026, 8:47pm

I’ve tried the following:

incus file remove collabora-code-container/etc/hosts

Then got:

$ incus start collabora-code-container
Error: Failed to run: /opt/incus/bin/incusd forklxc collabora-code-container /var/lib/incus/containers /run/incus/collabora-code-container/lxc.conf /var/log/incus/collabora-code-container: exit status 1
Try `incus info --show-log collabora-code-container` for more info

Checked the logs again:

$ incus info --show-log collabora-code-container
Name: collabora-code-container
Description:
Status: STOPPED
Type: container (application)
Architecture: aarch64
Created: 2025/08/29 23:55 UTC
Last Used: 2026/08/05 21:19 UTC

Log (lxc.log):

lxc collabora-code-container 20260805211905.768 ERROR    utils - ../src/lxc/utils.c:open_without_symlink:1196 - Too many levels of symbolic links - resolv.conf in /opt/incus/lib/lxc/rootfs/etc/resolv.conf was a symbolic link!
lxc collabora-code-container 20260805211905.768 ERROR    conf - ../src/lxc/conf.c:mount_entry:2101 - Too many levels of symbolic links - Failed to mount "/var/lib/incus/containers/collabora-code-container/network/resolv.conf" on "/opt/incus/lib/lxc/rootfs/etc/resolv.conf"
...

And then I’ve tried:

incus file remove collabora-code-container/etc/resolv.conf
incus start collabora-code-container

But the container does not really start:

$ incus info --show-log collabora-code-container
Name: collabora-code-container
Description:
Status: STOPPED
Type: container (application)
Architecture: aarch64
Created: 2025/08/29 23:55 UTC
Last Used: 2026/08/05 21:21 UTC

Log (lxc.log):

lxc 20260805212146.820 ERROR    af_unix - ../src/lxc/af_unix.c:lxc_abstract_unix_recv_fds_iov:218 - Connection reset by peer - Failed to receive response
lxc 20260805212146.820 ERROR    commands - ../src/lxc/commands.c:lxc_cmd_rsp_recv_fds:128 - Failed to receive file descriptors for command "get_init_pid"

Besides, isn’t there any option to disable the check on the container so I don’t have to remove files every time I update the container?

Thanks again!

pachulo · August 8, 2026, 9:09pm

Do you have any idea @stgraber ?

Thanks!

stgraber · August 9, 2026, 5:49am

Check incus console --show-log collabora-code-container.

A successful incus start but instance immediately goes to STOPPED usually suggests the init process / entrypoint immediately crashed/exited. So you need to look at its console to know what’s going on.

pachulo · August 9, 2026, 8:07am

This is what I get:

$ incus file remove collabora-code-container/etc/hosts
$ incus file remove collabora-code-container/etc/resolv.conf
$ incus start collabora-code-container
$ incus console --show-log collabora-code-container

$

By the way, this is how the /etc dir looks on the container:

$ sudo ls -lisah /var/lib/incus/storage-pools/default/containers/collabora-code-container/rootfs/etc
total 60K
518834 4.0K drwxr-xr-x  9 root root 4.0K Jul 23 20:43 .
518687 4.0K drwxr-xr-x 16 root root 4.0K Jan  1  1980 ..
519514 4.0K drwxr-xr-x  2 root root 4.0K Jul 23 20:43 apparmor.d
519509 4.0K drwxr-xr-x  3 root root 4.0K Jan  1  1980 apt
518843 4.0K drwxr-xr-x  3 root root 4.0K Jan  1  1980 ca-certificates
519515 4.0K drwxr-xr-x  2 root root 4.0K Jul 23 20:43 coolwsd
519503 4.0K lrwxrwxrwx  1 root root   71 Jan  1  1980 debconf.conf -> /nix/store/99kmbnjvwq0d5ypadfddr2pdhsazips1-image-root/etc/debconf.conf
518923 4.0K drwxr-xr-x  3 root root 4.0K Jul 23 20:43 fonts
519513 4.0K -rw-r--r--  1 root root  447 Jul 23 20:43 group
519512 4.0K lrwxrwxrwx  1 root root   64 Jan  1  1980 hosts -> /nix/store/99kmbnjvwq0d5ypadfddr2pdhsazips1-image-root/etc/hosts
518835 4.0K drwxr-xr-x  2 root root 4.0K Jan  1  1980 ld.so.conf.d
519496 4.0K lrwxrwxrwx  1 root root   72 Jan  1  1980 nsswitch.conf -> /nix/store/99kmbnjvwq0d5ypadfddr2pdhsazips1-image-root/etc/nsswitch.conf
519502 4.0K -rw-rw-r--  1 root root  885 Jul 23 20:43 passwd
519485 4.0K lrwxrwxrwx  1 root root   70 Jan  1  1980 resolv.conf -> /nix/store/99kmbnjvwq0d5ypadfddr2pdhsazips1-image-root/etc/resolv.conf
519504 4.0K drwxr-xr-x  4 root root 4.0K Jan  1  1980 ss

What are those /nix/store/99kmbnjvwq0d5ypadfddr2pdhsazips1-image-root paths?

stgraber · August 9, 2026, 11:08pm

Looks like their image is based on NixOS or on another distribution but with some stuff instead through Nix.

The fact that it doesn’t show anything on the console is slightly unfortunate as that makes debugging even more of a pain…

Do you have the same behavior if you do incus start collabora-code-container --console?