Hi Peff, Thanks for the help. $ GIT_CURL_VERBOSE=1 git clone http://git.site.domain.com/foo/gitrepo.git Cloning into 'gitrepo'... * Couldn't find host git.site.domain.com in the _netrc file; using defaults * timeout on name lookup is not supported * Trying 192.168.16.138... * TCP_NODELAY set * Connected to git.site.domain.com (192.168.16.138) port 80 (#0) > GET /foo/gitrepo.git/info/refs?service=git-upload-pack HTTP/1.1 Host: git.site.domain.com User-Agent: git/2.12.2.windows.2 Accept: */* Accept-Encoding: gzip Pragma: no-cache < HTTP/1.1 200 OK < Date: Fri, 28 Apr 2017 15:25:02 GMT < Server: Apache/2.4.6 (CentOS) PHP/5.4.16 < Last-Modified: Tue, 25 Apr 2017 18:11:35 GMT < ETag: "0-54e01a77ac500" < Accept-Ranges: bytes < Content-Length: 0 < Content-Type: text/plain; charset=UTF-8 < * Connection #0 to host git.site.domain.com left intact * Couldn't find host git.site.domain.com in the _netrc file; using defaults * Found bundle for host git.site.domain.com: 0x1cc9fc0 [can pipeline] * Re-using existing connection! (#0) with host git.site.domain.com * Connected to git.site.domain.com (192.168.16.138) port 80 (#0) > GET /foo/gitrepo.git/HEAD HTTP/1.1 Host: git.site.domain.com User-Agent: git/2.12.2.windows.2 Accept: */* Accept-Encoding: gzip Pragma: no-cache < HTTP/1.1 200 OK < Date: Fri, 28 Apr 2017 15:25:02 GMT < Server: Apache/2.4.6 (CentOS) PHP/5.4.16 < Last-Modified: Mon, 24 Apr 2017 20:51:42 GMT < ETag: "17-54defc6469818" < Accept-Ranges: bytes < Content-Length: 23 < * Connection #0 to host git.site.domain.com left intact warning: You appear to have cloned an empty repository. Content length is again 0. On Fri, Apr 28, 2017 at 11:20 AM, Jeff King <peff@xxxxxxxx> wrote: > On Fri, Apr 28, 2017 at 11:09:55AM -0400, Andrew Watson wrote: > >> Thanks for pointing me to git help http-backend. I confirmed the >> modules are loaded and the CGI environment variables. I've added >> "AcceptPathInfo On" to my httpd.conf just to be safe. >> >> I'm not sure what /info/refs is supposed to look like, but it is >> empty. Could that be the issue? > > No, that shouldn't matter. The on-disk file is used only for dumb-http > requests. In a working smart-http system, the info/refs request should > go to the CGI, which will generate the ref advertisement dynamically. > >> Do you see anything in my apache configuration that looks wrong? > > It looks reasonable to me, but I'm far from an expert on Apache config. > > When you clone, is it using smart-http there? Try using GIT_CURL_VERBOSE > to see what the response is to the initial /info/refs fetch when you > clone. > > -Peff