You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
magnusmanske edited this page Apr 17, 2012
·
2 revisions
Right now the ec2 is OpenSUSE and fairly minimalist, most tools don't even exist here.
Step one would be to install GIT:
yast2 -i git
Once git's been installed we need to install someone's key for Synbiota, and pull the repo.
CD into the web root, which on this ec2 instance is /srv/www/htdocs
The first time we run, we clone the repo
git clone git@github.com:Synbiota/GENtle2.git
Each time we want to update, we only need to CD into the root and run
git pull
Should we start creating branches and other things, the instructions here may become more complicated.
**todo: ** We need a dummy account as the private organization expects a privatekey, and my account is tied to... my account, which has other stuff in it. So, dummy count needs to get created (Connor should likely maintain this) with a key that gets installed. Since we can't do it that way, here's the other, messy way:
Use an SCP program to copy/replace files on web root.
GENtle2 config
GENtle2 needs the PHP CuRL and SSL packages. To install:
zypper in php5-curl
zypper in php5-openssl
edit /etc/php5/apache2/php.ini and uncomment line ";extension=php_openssl.dll"
apache2ctl -k restart
Each GENtle2 instance needs to have config.js.template copied to config.js, and edited to fit the installation (basically, Synbiota API URL). config.js will never be checked into git.