FS#12476 - [initscripts] start deamons in background improvment

Attached to Project: Arch Linux
Opened by Florian Pritz (bluewind) - Tuesday, 16 December 2008, 20:53 GMT
Last edited by Paul Mattal (paul) - Monday, 04 January 2010, 03:06 GMT
Task Type Feature Request
Category Packages: Core
Status Closed
Assigned To Aaron Griffin (phrakture)
Thomas Bächler (brain0)
Architecture All
Severity Low
Priority Normal
Reported Version None
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No
This task depends upon

Closed by  Paul Mattal (paul)
Monday, 04 January 2010, 03:06 GMT
Reason for closing:  Won't implement
Additional comments about closing:  Per Aaron: "What is unclear to me is what this gains us over the existing method. I get that we gain the ability to do something akin to DAEMONS=(foo @(bar baz)) with this change, but I don't think this is something we want to do."

Per patch author/submitter, he no longer uses this and it's untested.

Also, there have been no particularly compelling reasons stated (in the long life of this bug report) to implement it.
Comment by Glenn Matthys (RedShift) - Tuesday, 16 December 2008, 21:23 GMT
Question, have you actually measured how much time advantage this gives?
Comment by Florian Pritz (bluewind) - Tuesday, 16 December 2008, 21:27 GMT
On my desktop I switched to a static IP, but on the Laptop I'm still needing DHCP and there it saves me around 13 seconds which is quite much. Currently it boots in 20 secs iirc.
Comment by Gavin Bisesi (Daenyth) - Saturday, 21 March 2009, 14:35 GMT
Status?
Comment by Florian Pritz (bluewind) - Friday, 27 March 2009, 14:38 GMT
If you want to implement it, please pull
http://git.server-speed.net/cgit.cgi/initscripts/commit/?id=c8e56d9bd1ee30d06b76dcd0850e8d55ae554746
Otherwise I'd be happy to receive a status update.
Comment by Laszlo Papp (djszapi) - Saturday, 07 November 2009, 08:03 GMT
status ?
Comment by Thomas Bächler (brain0) - Saturday, 07 November 2009, 11:59 GMT
I am unsure whether to do this - it sounds rather complex. It has been sitting here for a while, so we should decide it though.
Comment by Aaron Griffin (phrakture) - Wednesday, 02 December 2009, 21:47 GMT
What is unclear to me is what this gains us over the existing method. I get that we gain the ability to do something akin to DAEMONS=(foo @(bar baz)) with this change, but I don't think this is something we want to do. This always comes down to the network daemon in the explanation. "I need to start FOO after I have an IP"... it just seems like there's much better ways to solve this problem. For one, can't netcfg do this?
Comment by Paul Mattal (paul) - Sunday, 06 December 2009, 19:47 GMT
I agree with Thomas that we should decide this. Let's get all arguments on either side in comments here by January's bug day, and we can decide around then.

I tried to follow the above link from Florian and got an empty repository. It would definitely be useful to see the implementation when considering the merits of this. I agree with Aaron that it sounds complicated, but maybe it is elegant-- it will probably depend somewhat on the implementation.

If we end up able to reasonably reconcile this with our KISS philosophy, it sounds like it could be beneficial.
Comment by Florian Pritz (bluewind) - Sunday, 06 December 2009, 20:13 GMT
Yeah, I removed that repo so here you have a patch for the current git HEAD.

Though I have to admit I don't use this anymore so it's untested. I switched to wicd on my laptop and static IP on the desktop and I actually don't reboot very often. The reason why I kicked it out is because for example OSS sometimes relinks the kernel modules and when it's backgrounded I 1. don't see the output and 2. don't have the device nodes when X starts -> all progs that need those can't start or can't output sound and I don't know when OSS finished relinking.
   0001-Add-support-for-starting... (2.5 KiB)

Loading...