"basically build an aufs stack and get out of there" That is the definition of what should happen in the *initrd*. The real init should be just that sysvinit with nice flexible shells scripts which setup things in the normal way, using fstab to mount the entries -using real mount.
Study the simple init process of any real slackware distro. Most distros have replaced all the sysvinit stuff with systemd which is much less flexible.
There should be no GUI stuff during bootup -if anything (for running 'Live') use the dialog utility. Keep it spare with any questions to the user during bootup -rather wait until the desktop is up and provide setup utils there for network, etc.
Rationalisation of init
Although usb and HDD booting is by far the most common approach I would highlight use of CD booting as the only real method by which a normal user can emulate a ROM boot - by which I mean that the code loaded into ram is pristine and unchanged each boot.gyro wrote:Yes, this won't help anyone booting from a cd, but maybe that's on the decline and booting from a writeable device like a usbstick or usbhd is becomming the external boot device of choice.
Both usb and HDD are writable media and can't be 100% relied on if the user wants identical boots each time, and also wants session changes dumped.
A closed CD is (as far as I am aware) unwritable and an extremely useful method to ensure reproducability of performance.
If we had some method of producing ROMs into which we could burn our favourite puppy we wouldnt need CDs but in the meantime i would love to see acceptance of optical media continued if possible.
@greengeek,
It seems to me that fewer and fewer new computers have optical drives.
But, I actually intended that paragraph to refer to the previous paragraph, i.e. the as yet unwritten utility that supports the /BOOT_SPECS file and so exposing boot parameters via a gui app. This utility wouild merge an updated version of the file into the initrd.gz of the running system. Much more complicated to do with a cd/dvd, than a writable device, and hence more likely to not get done.
gyro
It seems to me that fewer and fewer new computers have optical drives.
But, I actually intended that paragraph to refer to the previous paragraph, i.e. the as yet unwritten utility that supports the /BOOT_SPECS file and so exposing boot parameters via a gui app. This utility wouild merge an updated version of the file into the initrd.gz of the running system. Much more complicated to do with a cd/dvd, than a writable device, and hence more likely to not get done.
gyro
-
- Posts: 902
- Joined: Mon 22 Jun 2009, 01:36
- Location: Philadelphia, PA
the new init
gyro, please excuse; I'm just a user, not an expert.
Over the last several months, I've seen mentions of "the new init".
How can I tell whether a Puppy has this new init?
What are the main ways it works differently from older ones?
Thank you,
Sheldon Isaac
Oh, here's some info from the Puppy I'm in now, musher0's simplified Xenial:
Over the last several months, I've seen mentions of "the new init".
How can I tell whether a Puppy has this new init?
What are the main ways it works differently from older ones?
Thank you,
Sheldon Isaac
Oh, here's some info from the Puppy I'm in now, musher0's simplified Xenial:
Code: Select all
uname -a
Linux puppypc2261 4.1.2-EmSee-32-pae #1 SMP Wed Jul 15 12:39:34 BST 2015 i686 i686 i686 GNU/Linux
Code: Select all
ls -lat /sbin/init /initrd/pup_ro2/sbin/init /initrd/init
-rw-r--r-- 1 root root 54996 Dec 2 02:47 /initrd/init
-rwxr-xr-x 1 root root 10013 Oct 18 20:09 /initrd/pup_ro2/sbin/init
-rwxr-xr-x 1 root root 10013 Oct 18 20:09 /sbin/init
Dell E6410: BusterPup, BionicPup64, Xenial, etc
Intel DQ35JOE, Dell Vostro 430
Dell Inspiron, Acer Aspire One, EeePC 1018P
Intel DQ35JOE, Dell Vostro 430
Dell Inspiron, Acer Aspire One, EeePC 1018P
Re: the new init
The existence of files "/initrd/init" and "/initrd/tmp/HAVE_PARTS" would indicate that you have it.sheldonisaac wrote:How can I tell whether a Puppy has this new init?
Modern puppies built using woof-ce "testing" have it.
There are lots. If you really want to know then please read the preceeding posts in this thread.sheldonisaac wrote:What are the main ways it works differently from older ones?
But some are:
Internal algorithm is very different.
Has better support for internationalisation.
Extra-sfs's are now solely the responsibility of "sfs_load".
Better support for using boot parameters to specify the puppy sfs files to load.
Support for saving onto a partition other than the puppy frugal install partition.
gyro
Hello. I posted a thread asking for removal of flash drive boot code.
New flash drives ( USB & SD ) last about as long as H.D.s now.
http://www.murga-linux.com/puppy/viewto ... 258#980258
New flash drives ( USB & SD ) last about as long as H.D.s now.
http://www.murga-linux.com/puppy/viewto ... 258#980258
That's only if you have a USB 3.0 connection. Many people here run older hardware. For me even the old USB boot code doesn't wait long enough. See thread:sunburnt wrote:Hello. I posted a thread asking for removal of flash drive boot code.
New flash drives ( USB & SD ) last about as long as H.D.s now.
http://www.murga-linux.com/puppy/viewto ... 258#980258
Tahrpup Can't find Save folder on Large USB Hard Drive
Also note that if one connects many USB devices to a HUB they won't get USB 3.0 speed.