Hello Anaconda community,
Recently, I started noticing that my conda installation through apt-get is working locally, but failing on AWS eu-west-1 region. Following the installation procedure for Debian-based distros, we run:
apt-get update
apt-get install -y software-properties-common apt-transport-https ca-certificates apt-utils gnupg curl
add-apt-repository "deb [arch=amd64] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main"
curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor | apt-key add -
apt-get -
apt-get update
apt-get install -y conda
The installation on AWS fails on the second “apt-get update”
E: Failed to fetch https://repo.anaconda.com/pkgs/misc/debrepo/conda/dists/stable/main/binary-amd64/Packages.bz2 File has unexpected size (3287 != 3127). Mirror sync in progress? [IP: 104.16.131.3 443]
Based on this, I had a look over the InRelease file retrieved through my local machine and through the AWS instance.
Locally retrieved version of InRelease through a Cloudflare cache in Germany
Hash: SHA256
Origin: . stable
Label: . stable
Suite: stable
Codename: stable
Date: Tue, 19 Sep 2023 22:11:53 UTC
Architectures: amd64
Components: main
Description: Generated by aptly
MD5Sum:
...
0b6cada8bf459b31895205810e72dc18 3287 main/binary-amd64/Packages.bz2
...
AWS retrieved version of InRelease through a Cloudflare cache in Ireland
Hash: SHA256
Origin: . stable
Label: . stable
Suite: stable
Codename: stable
Date: Fri, 25 Aug 2023 15:22:57 UTC
Architectures: amd64
Components: main
Description: Generated by aptly
MD5Sum:
...
4e3f4fef2e27a41c7db375ada107276f 3127 main/binary-amd64/Packages.bz2
...
Notice the difference in the reported/expected file size for the Packages.bz2
Looking at the response headers the cache returns:
...
last-modified: Fri, 25 Aug 2023 17:10:15 GMT
cf-cache-status: HIT
cf-ray: 80ca0375d9fe956f-DUB
...
Looking at the other files, they appear to have been updated and match the ones I retrieve locally.
Therefore, I want to know if this can be solved from the conda maintainers or we have to wait arbitrarily until the cache is invalidated for the Dublin region. Let me know if I should report this somewhere else.