Puppy Wary 5.3 won't shutdown

Message

Tuber · #21 Post by **Tuber** » Fri 13 Jul 2012, 06:36

Okay, now I can confirm that it did shut-down properly when it was first installed. I just recreated my USB ext4 boot drive, and after setting the time zone and a few other basics I was able to power off shut-down.

EDIT: It breaks on the second boot. There is a message about doing something to the file system that will make it boot faster next time. After that it will no longer shut down.

Oh, and it may be relevant that I also have an Acer Aspire One Netbook. Model AO532h, Atom N450, Intel NM10 chipset, 2GB RAM, BIOS revision 1.26, Atheros wired and wireless network interface chips, Broadcom Bluetooth and CrystalHD decoder.

bigpup · #22 Post by **bigpup** » Fri 13 Jul 2012, 14:17

Tuber wrote:
bigpup wrote:Known fact. Some computers have a bios that was designed to expect a USB flash drive to be formatted VFAT. Drive manufactures format their USB flash drives VFAT by default.
Some computers have problems if VFAT is not used for the format of a USB flash drive.

Bootflash is designed specifically to setup a USB flash drive to work on any computer.
If that was the case, then we wouldn't even be able to boot our ext formatted flash drives. And I'd like to see proof of your "known fact" because I've never heard of such a thing in 30+ years of working in the computer industry. The BIOS only reads the bootloader from the MBR, then that takes over for the rest of the boot process.

Flash drive makers format vFAT by default because it can be read by virtually any device with a USB data port. It has nothing to do with making a bootable drive.

Well in early adoption of USB booting this was an issue.
Concerning Puppy presently, I will modify my statement by saying this:
To have a bootable flash drive (not booted from a bootloader on the hard drive)
If you keep the Flash drive Vfat format, the MBR on the flash drive will be setup to boot Vfat. Puppy can be installed with Universal installer and no changes need to be made to flash drive MBR.
If you use Gparted to format the flash drive, a Linux format (ext2,3,or4), the MBR will not be changed.
You will need to do something to the MBR to make it boot.
The easiest thing in Puppy is to run Grub4dos config and tell it to install a bootloader to the MBR of the flash drive.
(the MBR options in the Universal Installer, seem to only work with Vfat formats)

Tuber · #23 Post by **Tuber** » Sat 14 Jul 2012, 02:12

The "msdos" style boot sector has become a standard. Most popular file systems use the DOS type MBR. And that includes many of the Linux/Unix file systems. There are exceptions of course, like BSD UFS/GPT, MAC HFS and SUN Solaris stuff. But even some of the exceptions can use a DOS style boot sector for compatibility.

Gparted creates "msdos" boot sectors by default. When you go to the Device menu and create a partition table it will be "msdos" type unless you change it in advanced options. If you format the new partition ext4, the boot sector is still "msdos" type, with a Linux Volume Boot Record (provided it is a boot partition). The MBR simply points to the VBR so it's not OS specific. It is the VBR that loads and boots the file system (the volume boot record is not part of the master boot record, it is contained in the boot partition itself).

So yes, an ext boot drive still has a standard "msdos" master boot record.

And I put "msdos" in quotes because the MBR scheme was actually created by IBM.

Tuber · #24 Post by **Tuber** » Sat 14 Jul 2012, 08:00

And now we return to our scheduled topic...

Just tried Slacko, and that does shut-down correctly. So it looks like there is a problem with wary/racy. Slacko runs nicely on the little Netbook, so I'll just stick with that.

Monsie · #25 Post by **Monsie** » Sat 14 Jul 2012, 09:31

So it looks like there is a problem with wary/racy.

Maybe.... maybe not.

I run Wary Puppy... and I've run it in a few different scenarios. I would not call myself a master...yet...

but for the record, I have been running Wary Puppy 5.3 on a 2 gb usb drive that is formatted with ext 2, (for the past six weeks while I am trying another distro on my hard drive) and I haven't experienced any boot-up or shut down issues.

Here's some more specifics:

1999 Pentium 3 with usb 1.1 spec, no provision for usb boot from the bios
Running Wary 5.3 on single partitioned usb drive with help from boot CD
Used Universal Installer
left mbr alone... in other words did nothing to alter it
system files, sfs files, save file all in /mnt/home

From what I have read on-line, ext 2 is the best file system for a usb drive in terms of performance... but for everything you do, it seems there is usually a trade off. In this case, it could be that the file system is more prone to corruption than ext 3 or 4 since the new file systems have the journaling feature... but I am wondering if possibly the reason I am doing okay is because I used the Universal Installer (recommended), left the mbr alone, and did not partition the flash drive.

I'm not saying for sure that I have all the answers here, but am providing some food for thought.

Monsie

rjbrewer · #26 Post by **rjbrewer** » Sat 14 Jul 2012, 15:41

Monsie wrote:
So it looks like there is a problem with wary/racy.
Maybe.... maybe not.

Monsie

Their are numerous versions of Aspire netbooks with
different graphics and other hardware.

It's important to know the exact model number when
trying to diagnose problems.

Karl Godt · #27 Post by **Karl Godt** » Sat 14 Jul 2012, 19:47

There are some problems i encounterd with racy ( have not used wary until now ) with poweroff/reboot :

The code in /etc/rc.d/rc.shutdown kills users/programs that the " fuser " -binary detects are using it . The normal Puppy code does not filter anything in that case so it might be possible that fuser -k -m /mnt/sda# kills getty/mingetty, $PPID, init which would result in relogin or restart of X :

Do not delete /tmp/* at shutdown .

First investigations @ saluki

actual development

*

Another problem i have with rebooting my netbook using racy : the bios hangs while trying to detect drives .

Monsie · #28 Post by **Monsie** » Sun 15 Jul 2012, 08:08

Karl,

Thanks for those reference links.

As I understand the problem, the issues with rc.shutdown are only occuring in certain pupmodes, for instance pupmode=7 whereby one is booting directly from a usb drive and saving to that same drive also?

Are there other pupmodes that have this same issue?

Monsie

Karl Godt · #29 Post by **Karl Godt** » Sun 15 Jul 2012, 20:38

Monsie, no .

This is unrelated to PUPMODE==PuppyInstallationType .

I first thought it may have something to do with these since i had not such occurrence for a long time .

This seems to happen randomly that some Process-ID-Number obtained by `ps' or `fuser' commands come into the list of processes that could be closed/killed to be able to unmount a partition .

Mr Kauler uses a somewhat precaution in his scripts and seems no to like to upgrade coreutils for example which could break the syntax of his scripts, but fuser is the full binary (not busybox applet) and one of the newest . I think that probably older `fuser' might work better .

Monsie · #30 Post by **Monsie** » Tue 17 Jul 2012, 04:01

Well, I am glad that I asked those questions, because that is the way the issue appeared to me at first glance...

Thanks very much for the clarification, Karl.

Monsie

Tuber · #31 Post by **Tuber** » Thu 19 Jul 2012, 20:56

It could be a problem that doesn't affect most other computers.

My little Atom/NM10 Netbook has issues with another USB boot disk as well. Under the Parted Magic "Extras" menu you can boot the Super Grub, or Super Grub 2 disks, but on my Acer those 2 options cause a restart (they work fine on other computers).

Other than that, everything else I've tried has worked as expected. I currently have 6 different Linux distros installed for testing, and all were booted and installed from USB flash drives made with live CD/DVD images.

Tuber · #32 Post by **Tuber** » Thu 19 Jul 2012, 21:04

Monsie wrote:From what I have read on-line, ext 2 is the best file system for a usb drive in terms of performance...

Then you might want to read this.
http://www.linuxplanet.com/linuxplanet/tutorials/7208/1

greengeek · #33 Post by **greengeek** » Fri 20 Jul 2012, 00:10

Tuber wrote:EDIT: It breaks on the second boot. There is a message about doing something to the file system that will make it boot faster next time. After that it will no longer shut down.

Oh, and it may be relevant that I also have an Acer Aspire One Netbook. Model AO532h, Atom N450, Intel NM10 chipset, 2GB RAM, BIOS revision 1.26, Atheros wired and wireless network interface chips, Broadcom Bluetooth and CrystalHD decoder.

I had a series of strange problems on one of my Acer netbooks - it was an N450 based NAV51 (aka 350-21G16i as far as I can make out from the barcode sticker. MFG date 1009)

I had boot hangs, and also the "reboot to desktop" that you mention (Accompanied with a loss of access to the network hware...)

I tore my hair out for weeks trying to understand what was causing it and came to the following conclusions:

1) Several Puppies have scripts that alter the configuration on the second (and sometimes subsequent) boots. This means that things can seem "sorted" after a reboot, only to "fall back" to a nonfunctional state later on. I think this has something to do with filesystems not being mounted or dismounted correctly - at least judging by the comments I have seen from Karl and Shinobar regarding /tmp and fuser at shutdown time (I am totally unqualified to offer any useful comments about this - but it almost seems like some scripts "stack up" a list of changes to be done at shutdown or at next boot, but those scripts get sidetracked by something that happens to the filesystems and get into an unhealthy loop).
Maybe see Shinobars comments on this page:
http://www.murga-linux.com/puppy/viewto ... start=2805

2) Some BIOS's "remember" something (a hardware state or similar...) after the software shuts down, and for as long as the motherboard has power. (This is a hard one to explain - but on my Acer I got different results when I was booting from a USB stick depending on how recently I had powered my hardware off). Also - I have many times seen a BIOS behave differently depending on whether "wall power" has been removed. Some power supplies cause a pilot voltage to be supplied to the BIOS and/or usb buses until the wall plug is pulled. Some BIOS behave differently after you shut off to that "cold" state. I think of this type of BIOS as being a "live bios". I do wonder if battery-based devices like netbooks and laptops may have some strange interaction between the bios and ACPI depending on battery state etc - maybe the "poweroff" state is not truly a "poweroff" until the bios decides what is happening between the incoming wall power and battery charge circuit. I don't know.

3) I had been using the wrong method to write my mbr. In fact I was not rewriting the mbr - without realising it I was retaining the mbr that came on the usb stick and it contributed to the symptom I mentioned in (2) above.

Incidentally I had no joy with Bootflash (could have been my inexperience at fault...)

(FWIW some of the boot problems I had which illustrate the dynamic nature of the bios / mbr and usb bus are noted here:
http://www.murga-linux.com/puppy/viewtopic.php?t=78045
and here:
http://www.murga-linux.com/puppy/viewto ... start=2865

What this taught me is that there are mysteries that affect the specific actions occurring at boot time, especially on more modern hardware, and things are not as straightforward as they used to be when a bios was "dead". Boot actions and timings are affected by what the bios sees in terms of the readiness of it's peripherals, especially when they are on a usb bus (and particularly when that bus is driving SD cards as a boot source).

Probably none of this helps, but I thought I'd just throw it in the ring...

Monsie · #34 Post by **Monsie** » Fri 20 Jul 2012, 21:11

Tuber wrote:
Monsie wrote:From what I have read on-line, ext 2 is the best file system for a usb drive in terms of performance...
Then you might want to read this.
http://www.linuxplanet.com/linuxplanet/tutorials/7208/1

Thanks for that link; it was an interesting read. While the actual results varied between trials and between models or brand of flash drive, ext4 was the clear winner in terms of speed.

However, I should have explained what I meant by the term "performance". First of all, I was thinking in terms of Wary Puppy which is the basis for this thread. As such, I wasn't thinking of performance in terms of speed, but in the broader sense to include factors such as reliability and longevity. That said, I also noted that many things one does in computing involves some sort of trade off.... what one gains on one hand, one loses on the other as explained in this article http://www.ehow.com/info_12216958_puppy ... -ext3.html. On my old hardware, I chose ext2 in order to minimize the number of writes to my flash drive thereby extending its life. Speed was not a consideration for me, especially considering my usb 1.1 spec.

Your post prompted me to do some more research, and I found a solution... which is to format a usb drive with ext4 and disable the journaling feature as explained in this article: http://www.styryx.com/en/computers/oper ... -usb-flash For those who are running Racy and some of the other newer breeds on fairly new systems, then this would appear to be the way to go... but then this could really be the subject of another thread.

Monsie

Pelo · #35 Post by **Pelo** » Sun 02 Sep 2012, 16:37

Slacko 5.3.3 does not shut down (poweroff inop even with the terminal), neither via the menu , nor by the pup panel, rien de rien. jamais vu ça ! never seen that. Previous Slacko's : no problem at all.

(old)Puppy Linux Discussion Forum

(old)Puppy Linux Discussion Forum