| Author |
Message |
[MOB]Trip
Joined: 07 Oct 2009 Posts: 31 Location: same shit world u live in.
|
Posted: Wed 07 Oct 2009, 14:31 Post subject:
|
|
| mikeb wrote: | | Quote: | | define clean shutdown |
that sounds like you need 'acpi=force' as a boot parameter to save yer big toe.
This is about when puppy shuts down it does not unmount drives properly so leaves the dirty flag set so on a restart the file system has to be checked/cleaned up....it's doesn't usually mean any corruption but can be inconvenient.
mike |
and where would i put this acpi=force ehh?
and it appears to shutdown correctly...just doesnt shut the hell off afterwards.
_________________

|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Thu 08 Oct 2009, 00:31 Post subject:
|
|
Well if installed to the hard drive it would be on the kernel boot line in menu.lst which is in /boot...somewhere.....
something like
kernel /vmlinuz acpi=force .......
mike
|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Thu 08 Oct 2009, 13:07 Post subject:
|
|
Ok tries ataflash (and ideflash) on frugal puppy 4.12 but it remained in pupmode 12.
I did try on puppy 2.12 with ideflash and that did change to pupmode13.
One thing I just remembered is that as standard you cannot delete (hide) files from the original sfs in this mode. I did alter snapmergepuppy to correct this but there is still a problem if an original file is deleted and after a reboot it is restored... it remains hidden after the next reboot.
As mentioned before the puppy virtual space went from 256 MB (pup_save) to 1GB..(tmpfs in ram+swap) though of course space to save in remains at 256MB (or whatever the pup_save size is).
I really need to test this on puppy 4. I wonder if having /tmp in ram is affecting things...I have 512MB ram and 500MB swap....I will examine the init script.
mike
edit...of and dev_save is changed to pup_ro1...never knew why...so external links may not work.
was testing 4.12 retro and the init classes ataflash as fast media to will not use pupmode 13
|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Thu 08 Oct 2009, 19:22 Post subject:
|
|
Ok been experimenting with puppy 2.12....
I created an initrd.gz that performs an fsck on dev_save at boot.
basically the command line at the start of this topic does kill all apps but unfortunately this does not unlock the unionfs so dev_save remains mounted as rw regardless...pup_save does change to ro
Using ideflash although switching to mode 13 does not alter this behaviour.
So so far I have not found a means of performing a clean shutdown with a frugal install.
Now puppy 4 uses lazy unmount so next I must test this....
mike
nah 4.12 fails the acid test too....ideas welcome
mike
|
|
Back to top
|
|
 |
davids45

Joined: 26 Nov 2006 Posts: 580 Location: Chatswood, NSW
|
Posted: Thu 08 Oct 2009, 19:34 Post subject:
|
|
G'day [MOB]Trip
I fixed my shut-off problem by adding "power_off=1" to line 247 of 4.3's rc.sysinit (in /etc/rc.d).
There's a recent forum post that explains how this fix was found and works - please 'search' for this post as I am not able to give you the exact link from this computer I'm on at the moment.
David S.
|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Fri 09 Oct 2009, 12:28 Post subject:
|
|
Well I'm sure I'm going over old ground when it comes to frugal installs. Basically the unionfs is an intransient beastie.
Only other options would either be the suggestion discussed last year of loading the pup_save into ram too..more like how a multission puppy works, or a setup that I tried out where the full file system is in a 2fs/3fs image file which gets mounted by the initrd but unionfs is not used...either way a clean unmount should be pretty straightforward.
mike
|
|
Back to top
|
|
 |
gyro
Joined: 28 Oct 2008 Posts: 289 Location: Brisbane, Australia
|
Posted: Fri 09 Oct 2009, 12:51 Post subject:
|
|
mikeb,
Unlike 4.1.x and 4.2.x, 4.3 does set pupmode=13, if pmedia=ataflash is defined. The "init" script within initrd.gz has been changed in 4.3
gyro
|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Fri 09 Oct 2009, 13:48 Post subject:
|
|
| Quote: | | Unlike 4.1.x and 4.2.x, 4.3 does set pupmode=13, if pmedia=ataflash is defined. The "init" script within initrd.gz has been changed in 4.3 |
Thanks for the info..I guess it would be something like that.
From my brief, test since /initrd/mnt/dev_save is still locked up with unionfs there is stiil the unclean shutdown problem in mode 13
mike
|
|
Back to top
|
|
 |
gyro
Joined: 28 Oct 2008 Posts: 289 Location: Brisbane, Australia
|
Posted: Sat 10 Oct 2009, 12:12 Post subject:
|
|
| mikeb wrote: | | From my brief, test since /initrd/mnt/dev_save is still locked up with unionfs there is stiil the unclean shutdown problem in mode 13 | I've been running pup-430 as a frugal install on NTFS on my hard drive, in pupmode 13, using pmedia=ataflash, for a little while now, and each time I have done an fsck on pupsave.2fs from a different puppy frugal install, it comes up "clean".
gyro
|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Sat 10 Oct 2009, 12:42 Post subject:
|
|
| Quote: | | I've been running pup-430 as a frugal install on NTFS on my hard drive, in pupmode 13, using pmedia=ataflash, for a little while now, and each time I have done an fsck on pupsave.2fs from a different puppy frugal install, it comes up "clean". |
In that case 430 must be handling things in a better way than before....I assume without ataflash shutdowns are still dirty?
Well the topic was for full installs but without re-inventing the wheel it is good to look at the frugal behaviour.
Indeed if ataflash fixes the problem then it deserves it's own HOW TO
mike
|
|
Back to top
|
|
 |
gyro
Joined: 28 Oct 2008 Posts: 289 Location: Brisbane, Australia
|
Posted: Sat 10 Oct 2009, 17:19 Post subject:
|
|
| mikeb wrote: | | Well the topic was for full installs but without re-inventing the wheel it is good to look at the frugal behaviour. | I appreciate we've gone off topic, so this will be my last post in this topic.
| mikeb wrote: | | I assume without ataflash shutdowns are still dirty? | Yes.
| mikeb wrote: | Indeed if ataflash fixes the problem then it deserves it's own HOW TO | I's not sure I agree with "fixes", it's a workaround. And rc.shutdown still needs to be fixed for normal frugal installs.
gyro
|
|
Back to top
|
|
 |
mikeb

Joined: 23 Nov 2006 Posts: 4378
|
Posted: Sat 10 Oct 2009, 21:58 Post subject:
|
|
| Quote: | | I's not sure I agree with "fixes", it's a workaround. And rc.shutdown still needs to be fixed for normal frugal installs. |
Very true...I will get to grips with 4.30 and this subject...If I find a way to sort this for older puppies too that would be great.
Thanks for your input
regards
mike
|
|
Back to top
|
|
 |
|