Description of problem: $summary Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Ok so what do you mean by "really slow" ? Can we get a idea of the time it take for you, so we can measure improvement ? (as I was under the impression that gerrit was doing a git gc by itself) Also, to be sure, that's for the main git repo on gerrit, yes ?
git.gluster.org. Is that the same as gerrit? fedora 26 vm running on my desktop: %time git clone ssh://kalebskeithley.org/glusterfs.git dummy Cloning into 'dummy'... Enter passphrase for key '/home/kkeithle/.ssh/fedora_id_rsa': remote: Counting objects: 7908, done remote: Finding sources: 100% (1610/1610) remote: Total 150039 (delta 641), reused 149842 (delta 641) Receiving objects: 100% (150039/150039), 69.55 MiB | 114.00 KiB/s, done. Resolving deltas: 100% (95302/95302), done. 23.877u 2.510s 11:20.47 3.8% 0+0k 104+208696io 0pf+0w majority of the time is spent in Receiving objects. But on a fedora 25 vm also running on the same desktop Receiving objects is nearly instantaneous.
There's an automatic GC run every Sunday. A git clone takes about 30 seconds on my machine limited only by network speed. git.gluster.org is a CNAME to review.gluster.org. On ssh, it should do the same thing. On HTTP(S), it points to cgit.