Geo: Git clone with --depth 1 times out when the secondary is not up-to-date (#391980) · Issues · GitLab.org / GitLab · GitLab (original) (raw)

Skip to content

• • Why GitLab
  • Pricing
  • Contact Sales
  • Explore
• Why GitLab
• Pricing
• Contact Sales
• Explore
• Sign in
• Get free trial

Geo: Git clone with `--depth 1` times out when the secondary is not up-to-date

Customer ticket

A customer is noticing that git fetch from the Geo secondary randomly times out when the secondary is not up-to-date and the repo has to be proxied from the primary. They're on 15.8.3 and when the above condition is true, here's when it times out:

When --depth 1 is set during a git clone.

The customer reproduced the problem with the following steps:

• make a push of a 1MB file to a repository
• try to clone the repository by reaching the secondary geo instance and using. --depth argument
• the cli will get stuck

I was able to replicate this behavior with 15.8.1 with the following steps:

• Create a new empty project on the primary. Let it sync to the secondary.
• Push a large repo to the primary.
• Once uploaded and confirmed in the UI (and before enough time has passed to replicate the repo to the secondary), run git clone --depth 1 git@<PROJECT_ON_SECONDARY>. It hangs.

GitLabBook2:tests$ git clone --depth 1 git@SECONDARY_HOST:test-projects/rails-custom.git
Cloning into 'rails-custom'...
remote:
remote: This request to a Geo secondary node will be forwarded to the
remote: Geo primary node:
remote:
remote:   git@PRIMARY_HOST:test-projects/rails-custom.git
remote:
^C

I had to ctrl+C after about 10 minutes. Once the repository replicated to the secondary, the same clone command succeeded in just a few seconds.

Edited Nov 15, 2023 by Michael Kozono