Reloading httpd configs, instead of restart

[eDixon]

It would be nice if i could have nodeworx reload my apache configs instead of restarting the server. I do realize that every once in a while i will need to restart apache (such as after adding mod_deflate). But 99% or the time a simple reload will solve it (such as after modding the conf.d files).

This can be done by doing a simple:
# /sbin/service httpd reload

[timryberg]

If you've modified the configuration files then you're obviously already logged into the shell so what's the difference? Just type the command you gave before logging out and chech for errors. I would think that doing it from InterWorx would be doing double duty. (i.e. log out of the server, log into interworx reload if you get errors log back into the shell). I'm just wondering. Am I missing something here?

[lineman]

If you've modified the configuration files then you're obviously already logged into the shell so what's the difference?

That's what I was thinking. Maybe he's SFTPing or something.

[timryberg]

Yes, that's a possibility.

One of the "bugs" (if you want to call it that) in InterWorx is that if a service doesn't start it doesn't give you the error message that you get in the shell. If you fail to start httpd it always tells you what conf file and line the error is on, if this is done in the shell.

[IWorx-Chris]

One of the "bugs" (if you want to call it that) in InterWorx is that if a service doesn't start it doesn't give you the error message that you get in the shell. If you fail to start httpd it always tells you what conf file and line the error is on, if this is done in the shell.


The problem is that there is no standard to this output. if:


service <service name> status


could just return a sane return value (like 0 for 'up' and > 0 for 'down') it'd be fine, but since we can't get a good status from service < service name > it's hard to determine if a message is an error, a non-error or just some random status codes. :(.

Chris

[eDixon]

That is true, never expect me to see the obvious route ;)

I was just suggesting it because I find i reload about 20x more than restart, didnt occour to me that it not like i can edit those files w/o being sshed in anyway.

:).

[timryberg]

The problem is that there is no standard to this output. if:


service <service name> status


could just return a sane return value (like 0 for 'up' and > 0 for 'down') it'd be fine, but since we can't get a good status from service < service name > it's hard to determine if a message is an error, a non-error or just some random status codes. :(.

Chris

Thanks for the explanation, Chris. It's something I've always wondered. I suppose putting the output (whatever it was) could confuse some users.