Git Daemon won't stay up - Linux/Gitorious 64bit instance on Amazon EC2

I am running BitNami Gitorious 2.4.7-0 on a Micro EC2 instance. I have successfully updated config, hostname, hosts file. I can log into Gitorious and I have several projects/repositories set up, all pushing/pulling via SSH just fine. "Sometimes" I can access via git:// protocol from clients. I have to log on to Linux box and restart Gitorious frequently.

I set up a Jenkins server and attempted to pull via git:// protocol and I get a message:

Problem fetching from origin / origin - could be unavailable
...
stderr: fatal: unable to connect to subdomain.mydomain.com:
subdomain.mydomain.com[0: 10.xxx.xx.xxx]: errno=No error

Attempting to pull/clone in Git Bash from the same Jenkins machine results in:

stderr: fatal: unable to connect to subdomain.mydomain.com:
subdomain.mydomain.com[0: 10.xxx.xx.xxx]: errno=No error

Running

sudo ./ctlscript.sh restart gitorious

gives message "git-daemon not running" and then "git-daemon could not be started", but then magically some time later (5-10 minutes) I can access via git:// protocol and git-daemon has a valid PID. This only lasts a while though and soon I cannot clone via git:// anymore. I have to log on to Linux and restart gitorious (again giving messages that git-daemon is not running and then cannot be started). After a waiting a while I can clone via git://, then it crashes again... Lather, rinse, repeat.

Any ideas what could be causing the git-daemon to be crashing or stopping on its own?

I thought about upgrading to the latest, but I am not sure how to back up and restore all of my projects/repos into another instance.

Thanks!