(old)Puppy Linux Discussion Forum

jemimah wrote:I tested exclusively with ext2 filesystems, so that's probably it. Let me see if I can figure out how to fix it for ext3.

Edit: Wait.. Mstar you said you're using a .2fs, so it's probably not ext3 or at least not exclusively.

jemimah,

I wasn't sure whether Mstar was referring to the pupsave file itself as "ext2" or the filesystem that the pupsave was on.....

In my case the pupsave file itself is "ext2" and it resides on an "ext3" filesystem.

Thanks for checking this out,
s

When in the boot process do you get the error? Before it says "Copying Files to Ram", or after?

Also, are you using ext3 and grub on a USB drive? Or Syslinux on ext3 (is that even possible?). Can you try it with a Syslinux VFAT usb drive?

jemimah wrote:When in the boot process do you get the error? Before it says "Copying Files to Ram", or after?

Also, are you using ext3 and grub on a USB drive? Or Syslinux on ext3 (is that even possible?). Can you try it with a Syslinux VFAT usb drive?

jemimah,

Just ran it again and it gets to "copying files to ram......kb ram will be used" and then the "superblock ......" error happens.

Using grub on ext3 USB harddrive.

Thanks,
s

Do you get an actual number for how many KB of ram will be used? or just blank?

jemimah wrote:Do you get an actual number for how many KB of ram will be used? or just blank?

Just blank

So it's failing in the part where it calculates file size, then allocating a 0kb ramdisk. I wonder why. My internal drive is ext3 and it works fine.

jemimah wrote:So it's failing in the part where it calculates file size, then allocating a 0kb ramdisk. I wonder why. My internal drive is ext3 and it works fine.

Interesting.... only difference seems to be USB harddrive vs internal hd. Does the order of the kernel line parameters make any difference - I have the "ramboot=1" as the last.

I don't think so. Do you have anything for pmedia and pdev1? I'll post a new initrd with more debugging output when I get a chance to rebuild it.

jemimah wrote:I don't think so. Do you have anything for pmedia and pdev1? I'll post a new initrd with more debugging output when I get a chance to rebuild it.

pmedia=usbhd psubdir=pup-431

Maybe it's psubdir. I don't think I considered that...

jemimah wrote:Maybe it's psubdir. I don't think I considered that...

That would make sense, since the pupsave file loads just fine without the "ramboot=1" option.

Jemimah,

The frugal is on a FAT32 partition in its own subdirectory. The 2fs is the extension of the automatically generated pupsave file. I don't know if it's actually an ext2 file, but I presume it is. It's whatever Puppy creates.

Since I'm also using the psubdir command, that's a likely place to look for the problem. Here is the whole menu.lst entry:

#Rampup
title Rampup
rootnoverify (hd0,2)
kernel /Rampup/vmlinuz pemdia=satahd psubdir=Rampup ramboot=1
initrd /Rampup/initrd.gz
boot
#End Rampup

My installation quits in exactly the same way Seaside describes.

Ok, I'll check out the psubdir thing and see if I can fix it.

jemimah

(I have been eating red herrings again)
An OS started in ram
(do we have a networking booting Puppy)
would seem to be the most secure system
for instant shutdown.

Why might you want that?
Well let us say you have a Puppy droid
flying over a war zone dropping flower petals
(something I recently suggested to a former
defense minister)
Obviously you don't want your petal collection points
or flight path to fall into the hands of the militarily demented.

Droid down
No software.
The flowers are safe.

I've uploaded a new version that hopefully fixes this problem. I haven't fixed the shutdown script yet though. I'll look that that later this weekend.

Lobster, I believe I read there's a way to retrieve data from RAM after the machine has shutdown. But my understanding is you have only mere seconds to do this before it's to garbled to recover. So this solution is probably secure enough for your flower bomber Puppy.

jemimah wrote:I've uploaded a new version that hopefully fixes this problem. I haven't fixed the shutdown script yet though. I'll look that that later this weekend.

Lobster, I believe I read there's a way to retrieve data from RAM after the machine has shutdown. But my understanding is you have only mere seconds to do this before it's to garbled to recover. So this solution is probably secure enough for your flower bomber Puppy.

jemimah ,

Thank you very much for this update. I downloaded the new version and it worked flawlessly.

Now if I could only capture the thoughts that were in my head just after I shut down. No.... wait....that wouldn't be worth much

Thanks again,
s

Thank you Jemimah,

The pupsave is not saved on shutdown, however. If you are going to fix it, would it be possible to include the ability to re-size it as well as save it?

Just asking, I have no idea how it would be accomplished.

Thanks again for your effort and help.

I also do not get a pupsave file to save to the USB stick upon using the modified shutdown. I get an error message that says "Boot Device not found, file not saved". I did not take the usb stick out. Everything else seems to work when booting up from USB, meaning that the modified initrd says it detects the ramboot and then proceeds to bring the usb files (including a pupsave stashed on the usb from a previously unmodified initrd) to ram as it should. Unfortunately it somehow doesn't see the boot device when it shuts down...

Is there a way to manually copy the pupsave file from ram back to the usb stick when a real save is needed? I just wasn't sure is the /mnt version of pupsave is up to date. I tried to execute the save2usb script but it just says that the save is queued and the message never leaves the screen.

I've updated the shutdown script. It is now working for me, with or without psubdir set. Please try the new version.

Remember not to trust this script with your important data. Even if the code is perfectly correct, power outages, system crashes, and attempting to copy a mounted filesystem, will try to ruin your day at the worst possible moment.

A better way is not to use the shutdown script at all, and save any documents you are working on to the hard drive (outside the save file) or to a usb stick. If you need to install software or something just boot without the Ramboot option. That way you get a completely clean version of the OS every time you boot; like a remaster, but easier.

Putting everything totally to RAM limits a lot of HW tolerance Puppy is so proud of. I would limit the RAM-MANIA to those files and folders needed to speed-up the system. There's a lot of them, most of them very light, and a conclusive work of marking them as such and distinct them from others onRAM never really done. Perhaps we should have two to three initrds like initrd1.gz initrd2.gz... (one extra letter/number allowed) to let the user choose. I personally spend a lot of time wondering if separating/clouding some of RAM can open new horizons. Connections become more and more speedy, closer and closer to HD speeds abilities and reliability.