It's been a while since I've had to do this - and I may find the answer to this empirically myself, but just in case ...
Environment: TS / LS / OD 7.4.1.0 (potentially going to 7.4.1.2) on RHEL 6
Task: The customer does not want OpenDeploy running as root
Generally speaking, I know about the iwodnonroot CLT and have read the OD Admin Guide - but am a bit confused about two parts of the instructions:
1) After you run the iwodnonroot command, the documentation says:
"Start the OpenDeploy service as the non-root user ..." followed closely by: "NOTE: if you start the OpenDeploy service as root from the command line after configuring it to be run as non-root user, you must re-run the iwodnonroot command ...."
2) Following that is the section on configuring OpenDeploy to automatically start up as non-root after reboot. After you go through the whole process of creating the iwodserver_boot and iwodserver_boot_wrap and making all the symlinks and such, it says:
"To start and stop OpenDeploy after booting, use the following script:
/etc/init.d/iwodserver
rather than the script:
/etc/init.d/iwodserver_boot_wrap
..."
It doesn't state it clearly (to me) but since we would be using a sevice-account, where we don't have the password, and thus cannot login as them - we'd probably be using sudo, and if we run /etc/init.d/iwodserver via sudo (as root) - doesn't that defy the earlier "Note" and mean that we'd have to re-run the whole iwodnonroot process over again?
Or does it somehow magically set things up so after reboot, running /etc/init.d/iwodserver will actually perform the "su <non-root-user> -c ...." command??
I'm concerned because it's very easy to forget you have something set up as non-root and to run the commands you are used to running via sudo (as root) and not realize that you just broke something that was previously configured to run as non-root.