logentriescommunity / le_openshift Goto Github PK

If I stop the app using any command. And then start the app again using:

rhc app start -a {appname}

The monitor daemon never starts because the deploy action_hook is never called. I think you need to actually use the post_start action hook. I can explain more if you need help with it

It seems that in the post_deploy action hook you are looking for files with _log. to follow. This won't work for a lot of different cartridge types. It is too specific to apache logs. I think it would just be better to add *.log or add nothing and let the user add it in the ini file.

How to follow only error log?

FYI logentries team -- your current post_deploy bash file won't follow the Node logs, as they are saved as node.log -- filename doesn't include 'log'.

I've added the below to the bottom of my bash file (prior to start monitoring):

# Explicitly follow the Node and HAProxy logs
node=${OPENSHIFT_NODEJS_LOG_DIR}node.log
${OPENSHIFT_REPO_DIR}logentries/le follow "$node"
haproxy=${OPENSHIFT_HAPROXY_LOG_DIR}haproxy.log
${OPENSHIFT_REPO_DIR}logentries/le follow "$haproxy"

I just started a new PHP 5.4 cartrige on OpenShift Online today, and have been trying to get my logs sent to logentries.com. After following the installation instructions, no logs were being sent to logentries.com so I SSH'd into the app and ran le whoami, which said no logs were being followed. So I looked through the post_deploy script and noticed that env | grep OPENSHIFT_.*_LOG_DIR returned no resuts. The only environment variable that contained "LOG" was OPENSHIFT_LOG_DIR.
After I edited the post_deploy script, le whoami said it was following php.log