Really.... why

Asked by mapes

What is the point of this project. It seems to me to overly complicate daemon startup. I guess I'm just old school but, sysinit seemed to work perfectly. Heck even BSD's init is better than this.

Question information

Language:
English Edit question
Status:
Answered
For:
upstart Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Mike Buckley (michael-buckley) said :
#1

The goal I think was to speed up bootup. I'm sure Microso$t went through the same stuff with Windoze.

Revision history for this message
mapes (matt-bellizzi) said :
#2

A worthy goal to be sure. However I had an issue that just made me loose it. We had a new hire whom we told could install any distro she wanted on to the machine. She installed whatever is latest Ubunutu right now. Autofs was an upstart package NIS was not. (BTW it's quite clear Ubuntu cares not about nis users) Autofs was starting before nis and since we push maps via nis it broke. The only thing I wanted to do was have autofs start after nis. I found the documentation to be very lacking either from Ubuntu or from the upstart project page. What are valid start events? Is there a config file test or do I have reboot and try over and over to get it right. I found this to break already working functionally in sysinit (i.e I could just change S20autofs to S99autofs which is simple,elegant and error free). So what is the actual benefit of this that sysint does not provide? A engineer I know sums it up nicely

----quote
"that's why I hate linux. They keep dicking around with stuff and apps that compiled and ran just fine under version X suddenly don't under X+1. Gratuitous changes to APIs, replacing functionality that did not have problems. And when I look at what was done it looks like some project from intro to programming.

Everytime I've seen a big project decide to use linux they always end up having development tied to some weird version of one of the umpteen linux sources. They can't change because things won't work anymore. The environment is some snapshot in time that they're stuck in forever "
---endquote

SO this is a rant and you can delete it. I wasted some part of my life debugging this project or the Ubuntu implementation of this project and I'm basically done. Excuse me I have to go install BSD right now

Revision history for this message
Mike Buckley (michael-buckley) said :
#3

It has been frustrating for myself too as I have been dealing with a race condition with a /tmp filesystem encrypted using /dev/urandom as the key source, and is created and formatted at boot, then mounted by mountall. mountall fails to mount it about half the time and when it doesn't mount it, it locks the system. I spent about 20 hours coming up with a workaround which "seems" to work.

Can you help with this problem?

Provide an answer of your own, or ask mapes for more information if necessary.

To post a message you must log in.