Fixing Apache Proxy timeout issues
Apache server regularly timeouts while cloning large repositories and there is little advice on how to fix it. Most advice available concentrates on push, doesnt work for clone and doesnt fix the root issue with is Apache config.
Troubleshooting the issue, there were no problems with memory, processing, disk space. gunicorn continued creating the response even after the proxy had cut the connection. So attention needs to be on the proxy config.
Adding the following to the ProxyPass statement in your standard Apache config example (in the documentation) aligns the timeouts with Rhodecode and fixes the issue:
# Url to running RhodeCode instance. This is shown as `- URL:` when # running rccontrol status. ProxyPass / http://127.0.0.1:10002/ **timeout=7200 Keepalive=On** ProxyPassReverse / http://127.0.0.1:10002/
Could I recommend updating the relevant lines in your documentation?
I am using the community edition. Fully up to date. Its working well other than the issue with large repositories.
#1 Updated by Redmine Integration about 2 years ago
- Status changed from New to Resolved
Commit 64eeaebf1e97 by Marcin Kuzminski email@example.com on default branch changed this issue. https://internal-code.rhodecode.com/rhodecode-enterprise-ce/changeset/64eeaebf1e97629704e95fca09139fcb2ee2afa9
#2 Updated by Marcin Kuzminski [staff] about 2 years ago
Thanks for your contribution ! We added recommended settings into our example config.
Ps. RhodeCode is open-source, and we encourage people to add fixes themselves, simply register at code.rhodecode.com, and submit a Pull Request from your fork.
For example documentation is available here: https://code.rhodecode.com/rhodecode-enterprise-ce/files/default/docs