Re: lanuchd questions
Re: lanuchd questions
- Subject: Re: lanuchd questions
- From: Jason Coco <email@hidden>
- Date: Mon, 20 Oct 2008 15:33:41 -0400
On Oct 20, 2008, at 15:27 , Jordan K. Hubbard wrote: On Oct 20, 2008, at 7:04 AM, Dave Zarzycki wrote: A cheesy way to accomplish this goal would be leverage the launchd conditional keep alive logic and the fact that /var/run is guaranteed to be cleaned at boot. This would involve breaking each daemon out into a separate launchd job that is conditionally kept alive whenever /var/run/my-script-is-done exists. The script on the other hand, is also a launchd job, and is kept alive as long as /var/run/my-script-is-done does NOT exist. This script MUST "touch" /var/run/my-script-is-done to signal launchd to stop relaunching the script and begin relaunching the daemons.
You're right, that's pretty cheesy. :-)
Seems like we keep coming up against this "but I want these jobs to be sequential!" requirement again and again. "In the large" I also agree that sequential job launch is very fragile ("on demand" being so much more forgiving of unanticipated events and configuration changes) but I think there is still a legit need for small groups of things (4-5 items, max) to be able to depend on one another. Sounds like a job for the legendary "launchd helper"! :-)
- Jordan
P.S. If we don't want to write that helper, perhaps we should communicate our thoughts on it to a wider audience and see if someone else does!
I think you should! I had a need for something like that a month ago and I ended up writing my own...
J
|
Attachment:
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden