#Debian init
1 messages ยท Page 1 of 1 (latest)
Can you add --debug and share the output? Also, confirm what dagger version you're using.
Are you perhaps connected to a dagger runner in a different version?
Those logs are showing an older version of the python runtime being downloaded.
registry.dagger.io/engine:v0.10.3 this is the container running
Oh, sorry! You did show version was v0.10.3, but in helping someone else on v0.11.0 I got my wires crossed ๐
hehe, no worries. let me know if I can provide more info. I have not played enough with dagger to know where to look
Does the same error happen if it's dagger creating the new dir?
dagger init --sdk=python <new non-existing dir>
yes, the dir gets created but it is empty
With the same error?
the only docker-related settings I can think of is
Server Version: 24.0.5
...
Backing Filesystem: zfs
and a registry mirror
yes, exact same error
What version of Debian do you have?
lsb_release -a
No LSB modules are available.
Distributor ID: Debian
Description: Debian GNU/Linux 12 (bookworm)
Release: 12
Codename: bookworm
can see some errors in the dagger engine container logs as well
Yeah, see if something jumps out. I need to get a similar system to try to reproduce. Can you try another sdk? dagger init --sdk=go gopher
yeah, go works
uh, good to know
Latest 0.11.0 version has a whole new runtime for Python. Are you able to try the upgrade?
can I do this?
curl -L https://dl.dagger.io/dagger/install.sh | DAGGER_VERSION=0.11.0 sh?
Need to start cooking dinner here, but I'll give it a try and report back on Monday. Thank you for your help so far.
oh, saw 0.11.0 got released just an hour ago. I'll give that a try later tonight hopefully
๐
@mint herald, have you seen this error? Debian Bookworm, Docker 24.0.5, backed by ZFS.
No... that's a new one. I'll look quickly and open an issue if nothing glaringly obvious
It looks to me like it's due to WithoutDirectory("runtime") in https://github.com/dagger/dagger/blob/42a0f7a3d39221f3d869724de0c303fc8bb4a55b/sdk/python/runtime/main.go#L63.
Yeah it's definitely something involving that chunk of code, I'm just weirded out by the fact that we are internally trying to rm stuff from a cache mount (unlinkat /var/lib/dagger/runc-overlayfs/cachemounts) and that the error is "invalid argument" (one of the syscall errors involving something not existing would be less weird).
poking around though
also I can't repro it on my debian linux VM, which makes me slightly more suspicious ZFS could be involved somehow (no clue how though)
Same. That's what I was wondering as well.
@quasi tundra when you're around and have a minute could you provide the output of:
docker logs dagger-engine-9cf4d4ecdf5b646bsudo dmesg(this might be too big to send, in which casesudo dmesg | grep overlaywould still be helpful)
Also curious what cpu arch you have (uname -a); noticed a couple of weird things inside the engine image that are probably unrelated but would want to check the engine image for your arch too (I'm on arm64)