Cloning repositories should be performed in backgroud?

[Gustavo Leite de Mendonça Chaves]

I'm using RhodeCode 1.4.4 behind Apache. When I tried to clone an external Git repository I got kicked out by Apache's ProxyTimeout standard value of 300. It gave me a ProxyError page. However, I noticed that the cloning operation was still performing in the server until it completed and then I could see the repository in RC'd interface.

Wouldn't it be nice if RhodeCode kept the cloning in the background freeing its web interface? I mean, when one creates a repo cloning it from an external repo, instead of waiting, RhodeCode could give back immediatelly the new repository summary page with a clear message telling that the new repo cloning operation was still happening in the background. After the cloning operation finishes, RhodeCode would simply show the actual repository data.

The freeing of the web interface would give us a better web experience and it would be analogous with what GitHub and Bitbucket do. Moreover, it would spare us of having to mess with Apache's ProxyTimeout value.

What do you think?

[marcin]

Hi,

Do you talk about remote clones ? I totally agree that this should be turned into celery task, in the same way server side forking works. Than if you have celery enabled it will be done in a separate process and async. Can you open a ticket for this on bitbucket ? Thanks

[Gustavo Leite de Mendonça Chaves]

2012/10/22 marcin <mar...@python-blog.com>

Hi,

Do you talk about remote clones ?

Yes.

 

I totally agree that this should be turned into celery task, in the same way server side forking works. Than if you have celery enabled it will be done in a separate process and async. Can you open a ticket for this on bitbucket ? Thanks

I did it: https://bitbucket.org/marcinkuzminski/rhodecode/issue/613/perform-remote-clones-in-background

Thank you. 

--

Gustavo

[marcin]

This is now done (async if celery enabled)

cheers