usb Puppy Install (Wary 5.3.1)

Message

riccardo · #1 Post by **riccardo** » Mon 09 Apr 2012, 01:47

This is getting me down

I made a usb install (actually two of them) from a download of the latest Puppy Wary. On one ... a usb stick ... everything sees to work well. It starts and boots normally. The other ... an SD card of 2 Gig size ... seems to give me grief. On my desktop and thru a card reader it appeared to boot normally. However, when the card was placed into the socket on my laptop, the booting starts ... and gives the famous message about not finding the wary sfs file. It then drops me back into a useless shell.

I checked ... and the sfs file is there!! I redid the installation with the same result.

The installations were done on a P3 computer running a live puppy wary ... and after the sd-card was formatted FAT32 by the puppy installer. I've checked the syslinux.cfg file on the sd-card ... and it is exactly the same as for the usb stick.

Most irritating!! I want to use the SD card because it is faster and (being stored inside the computer) is less prone to being 'lost' or damaged by the elderly (all thumbs) operator!

riccardo

rcrsn51 · #2 Post by **rcrsn51** » Mon 09 Apr 2012, 01:55

There has been a lot of discussion recently about problems with booting off SD cards. Basically, the SD card is slow to initialize, so Puppy doesn't detect it at bootup, and the pupxxx.sfs file is not found.

This will eventually be fixed. But for the time being, try starting your machine with a non-bootable flash drive plugged in. Its presence may slow down the boot process.

Also, delete the "pmedia=usbflash" argument from the syslinux.cfg file on your SD card.

riccardo · #3 Post by **riccardo** » Mon 09 Apr 2012, 02:09

Thanks for your reply my friend. It is (sort of) satisfying to hear that others have experienced similar problems.

1. I deleted the pmedia=usb section on my syslinux.cfg file. NO change in the result. If the problems was there, then more will have to be done!!

2. It surprises me the the system works with a usb stick, because I feel these are on the whole slower than the sd-card!! Maybe they start up/initialise more quickly.

3. I wonder if others have had greater success by reverting to an older version of puppy. My hardware doesn't really need Wary ... and maybe I will have success with an older Puppy ... say Lupu version 5.2.8??

4. Maybe what is wanted is a small piece of code which amounts to a wait instruction ... so that booting starts and delays a couple of seconds until the sd-card is initialised??

Thank you again for your thoughts...

riccardo

riccardo · #4 Post by **riccardo** » Mon 09 Apr 2012, 04:05

While on a walk this Easter Monday, I had a small brainwave. Upon my return home I put the SD-card into the card reader and then plugged into the usb port on my netook. Rebooted and it just worked! So, my feeling is that the puppy linux to sd-card installation has worked, but that either the startup script is too impatient OR the SD-card slot on my laptop is faulty.

Data can be read off a card in the slot, so I'm leaning towards the former explanation.

As an aside, I watched the screen carefully as the startup looked for the sfs file ... and no less than 6 white dots appeared while it was looking, and then 6 more red dots. This suggests to me that my SD-card is slow to start ... whatever its data rate is!!

I could survive with the netbook in this configuration, but then why use the sd-card???

Still looking to cure my impatient startup scripts...

Riccardo

Monsie · #5 Post by **Monsie** » Mon 09 Apr 2012, 04:39

Hi riccardo,

First, I'd like to clarify which version of Puppy you are using... your subject heading says Wary 5.3.1 which should be Wary 5.3 unless... it turns out you are actually running Slacko which is at 5.3.1

Please have a look at this recent thread which addresses your issue about the save file not being found on the SD card during the boot process. There is one notable difference in this scenario, and that is that the user is booting off the Puppy CD, however, this may not matter if Puppy can find your save file during boot-up.

Hope this helps,
Monsie

riccardo · #6 Post by **riccardo** » Mon 09 Apr 2012, 05:06

Thanks for your reply. You are correct: I had the title wrong. I have been using Wary 5.3

Did you see my later posting re. the system works in my card reader. It does keep me interested. I shall follow up the leads in your email.

Riccardo

riccardo · #7 Post by **riccardo** » Mon 09 Apr 2012, 05:30

Monsie et al.

Tried both of the suggestions without success.

Telling the booting process where the sfs file was not helpful. It had previously gone to the root directory to quickly find both the kernel and the initrd files. That's what is puzzling!! The sfs file is in the same place as these two!!

Going back to putting the card into the card reader produced the working result ... even though it does take additional time for the sfs card to be found.

Thanks again for your inputs...

Riccardo

BarryK · #8 Post by **BarryK** » Mon 09 Apr 2012, 08:15

Riccardo,
Would you be interested in conducting some experiments?

I did implement what I thought was a fix for SD slowness problem, and I tested it on my Acer 3681WXMI, and it worked. However, maybe for others it needs more delay.

This is my blog report about the fix:

http://bkhome.org/blog/?viewDetailed=02771

The script that I modified in initrd.gz is /sbin/wait4usb. You could modify that, try inserting a delay. Perhaps try this first (with SD card in card slot, not in usb adapter):

Test 1
In the 'syslinux.conf' on the SD card, insert "pfix=copy,rdsh" as a kernel boot parameter. Um, let's see, it will look like this:

default vmlinuz initrd=initrd.gz pfix=copy,rdsh

Boot with SD card inserted. This will drop you to a shell inside the initramfs. Type this:

# /sbin/probepart_init

...this will confirm whether Puppy is able to see your mmc (SD) card or not. If so, then proceed to Test 2. if not, then probably a needed kernel module is missing from the initrd.gz.

Test 2
Do you have a USB flash pen drive that is just memory, not bootable (that is, does not have any Linux installed on it)? If so, plug that in when laptop is off. Turn on laptop, booting from SD card. The probing for the USB drive will introduce extra delay, that might be enough for the SD card to get recognised. Let us know,

Remove Flash pen drive.

Test 3
Modify the initrd.gz. See my instructions for doing that here:

http://bkhome.org/blog/?viewDetailed=02787

...in future, this will be a standard feature in puppy.

Edit sbin/wait4usb, insert a big sleep here:

Code: Select all

sleep 10
#120330 probe for mmc drives...
MMCDRVS=""
if [ "`grep '^mmc' /tmp/ALLDRVS0`" = "" ];then  #find out if already available.
 [ "`lsmod | grep '^mmc'`" != "" ] && MMCDRVS="`find /sys/block -maxdepth 1 -name 'mmc*' | tr '\n' ' '`"
 [ "$MMCDRVS" = " " ] && MMCDRVS=""
fi

echo -n "$ALLUSBDRVS" > /tmp/flag-usb-ready
[ "$MMCDRVS" ] && echo -n " ${MMCDRVS}" >> /tmp/flag-usb-ready

###end###

...if that fixes it, find the smallest sleep that works.

Thanks for your help!

riccardo · #9 Post by **riccardo** » Mon 09 Apr 2012, 09:34

Barry K.

Thanks enormously for your contribution. I shall try out the experiments tomorrow ... and let you know the results. That's along the lines of my muses! Great stuff.

Riccardo

Karl Godt · #10 Post by **Karl Godt** » Mon 09 Apr 2012, 17:42

I think that there is more behind this . It seems that some usb readers can block the accces by setting a bit . I have experienced that with one USB-SD-Card-Reader and one MP3-Player . The MP3 player boots with Grub4dos fine, but then seems to block the access by the bios .
This only applies for cold poweron of the PC .
If the PC is rebooted, the MP3-Player isn't recognized by the BIOS at all and wont show up in the Boot Menu of the BIOS .

From the rdsh i was able to boot to desktop by mounting proc (mount -t proc none /proc) to give switch_root access to the PIDS, running ./init again (more than once) and run " exec switch " .

rc.sysinit complained /proc (and /sys ) already mounted but did not hang and booted to desktop .

And further note : my sd cards and mp3 players and micro-sd-cards show off as (usb) /dev/sdb* -drives, not as /dev/mmc*

if [ "`grep '^mmc' /tmp/ALLDRVS0`" = "" ];then #find out if already available.
[ "`lsmod | grep '^mmc'`" != "" ] && MMCDRVS0="`find /sys/block -maxdepth 1 -name 'mmc*'`"
fi

riccardo · #11 Post by **riccardo** » Tue 10 Apr 2012, 03:14

Barry

Thanks again for your suggestions. I edited the syslinux.cfg file as you suggest ... taking out the line 'append initrd=initrd.gx pfix=copy' and replacing it with 'default vmlinuz initrd=initrd.gz pfix=copy,rdsh'. Then I rebooted with the sd-card inserted. The booting stopped at the usual place ... looking for the sfs file ...and dropped me back to the shell. Undaunted, I then ran the command ' /sbin/probepart_init ', and the resulting output included /dev/sdb1. According to your notes, I think this means that the sd-card is being seen by the kernel.

While looking about for another sd-card on the bench, I came across another 2 GB one. So, I inserted it in the slot and rebooted. To my amazement, it was an older copy of Lupu 5.2.5! ... And it booted smoothly!!!!! I think I used a different boot-loader when I installed it ... as the files on the sd-card include ldlinux.sys and the conf file is extlinux.conf. The contents of this file seems very very similar to my original syslinux.cfg! The hardware is the same ...an sd-card from 'Scandisk'. For the moment ... it has set me wondering if I would do better to go back and try Lupu 5.2.8 ... with a different boot loader.

What do you think about this??

Riccardo

mavrothal · #12 Post by **mavrothal** » Tue 10 Apr 2012, 11:09

Not all cards are "created equal" even from the same manufacturer.

You may want to try "PDEV1=mmcblk0p1" as a boot code with the "bad" card or instal the wary files in the "good" Lupu525 card and try.

Also, are both cards VFAT formatted?
I still do

Code: Select all

sed -i "s/vfat)/vfat) \\n   modprobe vfat/" init

for the XO builds (that are usually in an SDcard), just to be on the safe side.

mahaju · #13 Post by **mahaju** » Tue 10 Apr 2012, 12:44

I have done a frugal install on an SD-card and use it to boot to puppy 5.1.1 through the built in SD-card reader on my Dell laptop
I had booted the same laptop using the puppy CD and did a frugal install on the sd-card from the CD boot
In my case there was no problem at all
My laptop by default goes to Windows 7 on the hard drive, so when I have to boot puppy I first insert the sd-card, turn on the computer, bring up the boot menu and then select the sd-card for booting
I don't know for sure, but perhaps there are some bios settings that we can alter, related to the sd-card?

Karl Godt · #14 Post by **Karl Godt** » Tue 10 Apr 2012, 14:20

So, I inserted it in the slot and rebooted.

I have a Dell that does not recognize at a cold poweron an attached USB IDE/SATA converter , but when i reboot, after one kernel had booted, the external HDD is shown in the boot menu .

There must be something different in warm and cold starts at some boards .

capoverde · #15 Post by **capoverde** » Tue 08 May 2012, 10:09

This kind of thing has been puzzling me for some time now: the same SD card, with one or another Puppy version installed, boots fine on one netbook/laptop and doesn't boot on another. Sometimes it works fine when inserted in a USB card reader and doesn't even begin booting from the onboard SD slot. On some machines it does boot if the install was done as a USB-HD, on others it doesn't; etc. etc.

It appears that there are several different reasons for this irregular behavior, some depending on the card, some on the BIOS and some on the hardware being used -- which makes things difficult. Haven't tried to sort it out, just got to boot each box by plain ol' trial-and-error... All suggestions are highly appreciated.

(old)Puppy Linux Discussion Forum

(old)Puppy Linux Discussion Forum