dizcza/docker-hashcat

error 0002

Closed this issue · 3 comments

erenmlbb commented

i tried pulling it on docker and got this

docker run --gpus all -it dizcza/docker-hashcat /bin/bash
Using default tag: latest
latest: Pulling from dizcza/docker-hashcat
76769433fd8a: Pulling fs layer
73c18e36c3d4: Pull complete
ba1ec7a08b4c: Pull complete
e321dcf9098b: Pull complete
909b2587c444: Pull complete
835d627a028a: Pull complete
46a0313239a3: Pull complete
a84a076e227e: Pull complete
2c05008dc145: Pull complete
a5b22467f1c7: Pull complete
26f29e1250b4: Pull complete
0d2189f7abf6: Pull complete
58247411b337: Pull complete
c2298d521b5f: Pull complete
53ca54e101fd: Pull complete
90eff9de5e71: Pull complete
4f4fb700ef54: Pull complete
642667af9803: Pull complete
f04a9d11ac93: Pull complete
f6294fff0407: Pull complete
22aaee882a9f: Pull complete
d15f0844fb52: Pull complete
55e37b51b969: Pull complete
08da2354cb07: Pull complete
Digest: sha256:f8281364588bc5f443049f8d7718b3c4747339d45b8a300f20d1329501b8f9e5
Status: Downloaded newer image for dizcza/docker-hashcat:latest
docker.io/dizcza/docker-hashcat:latest
docker: Error response from daemon: could not select device driver "" with capabilities: [[gpu]].
ERRO[0002] error waiting for container: context canceled

dizcza commented

Well, it means that docker does not see / have no permission to use gpus.
What's your nvidia-smi?
Make sure that your user is in the docker group: running groups should output docker among other things.
After you make sure that nvidia-smi behaves correctly and you're in the docker group, run nvidia-smi within an image as described https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/sample-workload.html

If that works for you and you still see the issue specifically for my docker images, let me know.

erenmlbb commented

Im actually using amd gpu on parrotos 6.0 my kernel is 6.5 but it seems that drivers from official amd site is only up to 6.3.6

dizcza commented

Sorry, I cannot test AMD gpus - I don't have any. The latest tag should work for AMD gpus because in this docker image there is an OpenCL backend. I'm not sure though about the --gpus all flag for running a docker container on an AMD GPU - I know this flag should be used with nvidia gpus, try removing this flag or investigate how to pass AMD gpus to a docker image.

In either case, you can try the pocl tag, I don't know if POCL supports AMD gpus - maybe.