apt update failed on ubuntu 18.04

[laohyx]

Hi team,

I know that it is not related with the docker. But I cannot find where to report this question.

This problem seems like #619

---

1. Issue or feature description

When running apt update with Nvidia source is configured, Packages.gz mismatchs with Release file's content.

2. Steps to reproduce the issue

$ wget https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/Packages.gz
--2020-06-21 10:38:34--  https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/Packages.gz
Resolving developer.download.nvidia.cn (developer.download.nvidia.cn)... 183.131.64.232, 183.146.27.25, 183.131.64.231
Connecting to developer.download.nvidia.cn (developer.download.nvidia.cn)|183.131.64.232|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 32924 (32K) [application/x-gzip]
Saving to: ‘Packages.gz’

Packages.gz         100%[==================>]  32.15K  --.-KB/s    in 0.03s   

2020-06-21 10:38:36 (1020 KB/s) - ‘Packages.gz’ saved [32924/32924]

$ md5sum Packages.gz 
6a7d71edd4c29ad8c382d9cf169ded55  Packages.gz
$ sha1sum Packages.gz 
09bc6f217ed3bf63c7d65275a6285b44c558de17  Packages.gz
$ curl https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/Packages.gz.md5
33c6dadca2bfdb310ba05f654b9931ca Packages.gz
$ curl https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/Release
Origin: NVIDIA
Label: NVIDIA CUDA
Architecture: x86_64
Date: Wed, 17 Jun 2020 21:47:53 +0000
MD5Sum:
 755245ae201cbb29a79f11b719308204           150181 Packages
 33c6dadca2bfdb310ba05f654b9931ca            38727 Packages.gz
SHA1:
 2b72af3d50ae635917c1673d45b9431d5de5aa44           150181 Packages
 ef946791bf29f350ad12d1a4e68d578ea2d89cd1            38727 Packages.gz
SHA256:
 d6358a1363e130724ed5a93d7143f6f4e5f26a41cc7e89e305c4171984e68610           150181 Packages
 95fed63a913caa0e383c61a518cbd554cfd8b593251eca07f004b8afd894d959            38727 Packages.gz

You can see:

name	described in Release file	What I got
size of Packages.gz	38727	32924
md5	33c6dadca2bfdb310ba05f654b9931ca	6a7d71edd4c29ad8c382d9cf169ded55
sha1	ef946791bf29f350ad12d1a4e68d578ea2d89cd1	09bc6f217ed3bf63c7d65275a6285b44c558de17

Is that a CDN issue? My location is Shanghai, China. But I tried again after 10 hours, and the problem still exists.

Thanks.

3. Information to attach (optional if deemed irrelevant)

• Some nvidia-container information: nvidia-container-cli -k -d /dev/tty info
• Kernel version from uname -a
• Any relevant kernel output lines from dmesg
• Driver information from nvidia-smi -a
• Docker version from docker version
• NVIDIA packages version from dpkg -l '*nvidia*' or rpm -qa '*nvidia*'
• NVIDIA container library version from nvidia-container-cli -V
• NVIDIA container library logs (see troubleshooting)
• Docker command, image and tag used

Thanks.

[nvjmayo]

also reported on #1328

[didibaba]

Also got the same problem.

[vamosdalian]

why always skip to another question?
who can answer the question?

[lukezhqin]

would it possible to use proxy?

[vitahlin]

Do you resolve this problem ?
Who can answer this question?

[klueska]

This is not an issue for nvidia-docker itself.

All issues related to images running on top of nvidia-docker should be directed here:
https://forums.developer.nvidia.com/c/accelerated-computing/cuda/cuda-setup-and-installation/8

[vitahlin]

You can try like this:

1. Get IP of developer.download.nvidia.com

ping -c 2  developer.download.nvidia.com
PING hnd.x.uxengine.net (128.1.83.18): 56 data bytes
Request timeout for icmp_seq 0

--- hnd.x.uxengine.net ping statistics ---
2 packets transmitted, 0 packets received, 100.0% packet loss

2. Modify the file to redirect the domaindeveloper.download.nvidia.cn to IP 128.1.83.18：

vi /etc/hosts

# Add this
128.1.83.18 developer.download.nvidia.cn