Puppy Linux Discussion Forum

tallboy · Joined: 21 Sep 2010 Posts: 343 Location: Oslo, Norway

Flash: Smart! Cool

Update: Small correction, the CD going into kernel panic contained only one small save.

I did a new test, and everything froze using 'puppy pfix=1', but NOT when using 'puppy pfix=ram' ! Question

Tallboy.

Flash · Posted: Wed 15 Jun 2011, 23:37 Post subject:

The same thing happened to me when I tried to boot a multisession DVD which had only one session saved on it. You'd think Puppy would boot as if there were no saved sessions at all when you use pfix=1 with only one session saved on the DVD, but evidently not....

So I switched to a multisession DVD with many saved sessions on it, booting with puppy pfix=2 (choosing the 2 at random), then saving to the DVD and rebooting. What I found after mounting the DVD was a new hidden file on the DVD named .badfolders inside the latest saved session on the DVD, and a new hidden file named /.badfolders in Puppy's root directory. Inside .badfolders are the two sessions I blacklisted with the puppy pfix=2 boot option. I figured that if I replaced the two sessions in /.badfolders with any of the valid session(s) on the DVD and then shut down saving to the DVD, Puppy would skip the sessions I put in .badfolders when it boots, and that did indeed happen.

gcmartin · Joined: 14 Oct 2005 Posts: 2627 Location: Earth

tallboy · Joined: 21 Sep 2010 Posts: 343 Location: Oslo, Norway

Flash, what happens if you set bootmessage pfix=1 to the DVD with many saves on it?

It is a pity the construction don't work properly, and especially without having a dialog window informing of the situation, and asking if you want a .badfolders file written. I thought that only the boot process noted that file(s) should not be loaded, and that it was strictly for the running session only. I can see why we need such a function, to avoid loading the last saved file that may contain settings for another machine with a different cpu. It will be difficult to understand for someone not being aware of the fact that all modifications and settings in the file marked bad are gone, just because he/she saves again.

Isn't there a file describing which files that are saved in a normal session save? Then there is possibility to omit the .badfolders. That action also need a dialog window.

I still don't understand the reason for 'Kernel panic - not syncing: Attempt to kill init!'

Tallboy.

Flash · Posted: Fri 17 Jun 2011, 00:44 Post subject:

8-bit · Joined: 03 Apr 2007 Posts: 3012 Location: Oregon

In my case, first, I am using a DVD +RW media and going multisession with Puppy Lucid 5.2.0.
I got most things settings done on the first boot with Pfix=ram and then doing a SAVE.
I then went back after finding my wifi was not supported to the extent of not showing up.
I used ndiswrapper and got everything set up and saved before saving back to the DVD.
But....
Upon a reboot, later after adding some SFS files, my network settings had disappeared.
I noticed the order the sessions were loaded was latest first and then older sessions in order.
So are the older sessions overwriting the network settings when loaded?
I currently have 4 sessions on the DVD so I assume to use the most current one and ignore the older 3 I would boot with "pfix=3" to get the others blacklisted.

And FYI, if one installs the latest version of sfs_load a right-click on an SFS file will offer to load it with a warning of it being experimental.
But I was able to load SFS files that way.
SFS_Load-on the fly will not see SFS files on the DVD though that I have in my /root/my-documents directory.
I have 4gigs of memory on the laptop so putting them in the session did not bother me.

I still have to try putting the SFS files in the root of the DVD not to be confused with /root to see if it messes up booting ability.

What I do not understand though is to why the sessions are loaded in reverse order of the dates.
If they were instead loaded in earliest to latest order, it seems like the changes one made to network settings would stick.
Currently, I have to go through the whole network setup process every time I boot the DVD.

Jasper · Joined: 25 Apr 2010 Posts: 869 Location: England

Hi 8-bit,

Your assumption is wrong, "puppy pfix=3" loses the latest 3
and retains the original of your four sessions. You can
confirm this by pressing F2 on boot and reading.

BarryK, and others, understand the multisession boot
order - and it works. Do not worry about the "experi-
mental" note either.

My regards

Flash · Posted: Sun 09 Dec 2012, 09:21 Post subject:

I wondered why multisession Puppy loads the sessions in counterintuitive order when it boots. A few years ago I actually modified a Puppy so it loaded the saved sessions first-to-latest. It made no difference. I posted my findings somewhere in this forum. Then someone who knows more than I do said that the reverse order was done on purpose. The upshot is, the way multisession Puppy combines them when it boots is indifferent to the order sessions are loaded.

8-bit · Joined: 03 Apr 2007 Posts: 3012 Location: Oregon

You have lost me on that one then.
That is to say if I have 4 sessions with say first, second, third, and fourth that using pfix=3 will blacklist second,third,and fourth using only first?
It would seem that the pfix=3 should instead use the fourth one and ignore first, second, and third.
IOW, how do I know that the earlier sessions when loaded are not overwriting my network settings that are contained in both the third and forth, but not the first and second?
As I had previously stated, I am using ndiswrapper, blacklisting eth0, and using a windows driver for my wifi.
It does work when I set it up. But the network settings I made are not being loaded from the fourth session.
I am shown as not having a network connection on booting and eth0 shows up as the only wifi interface.
Having to set up my network settings each time I boot is not something I should have to do!

And how, would one use just the fourth session while ignoring the first three in that case?

Flash · Posted: Sun 09 Dec 2012, 14:34 Post subject:

Sorry, I was only talking about the order in which multisession Puppy loads the sessions when it boots without using the pfix option.

gcmartin · Joined: 14 Oct 2005 Posts: 2627 Location: Earth

Linux does not approach the use of the copy command in the same vane as Microsoft.

This is why it starts with the very last Savefile first; thus anything from older saved files will NOT overlay an existing (potentially newer) file in any of the folders.

In Microsoft, a copy operation would be done starting with the oldest to achieve the very same thing; thus it would overlay something from an older folder with the contents of the newer.

Hope you understanding the differences in copy operation across the OSes.

Here to help

Q5sys · Joined: 11 Dec 2008 Posts: 843

gcmartin · Joined: 14 Oct 2005 Posts: 2627 Location: Earth