Pup214R v1.00 - Puppy Linux 2.14 Revisited, is now available

News, happenings
Message
Author
User avatar
Dougal
Posts: 2502
Joined: Wed 19 Oct 2005, 13:06
Location: Hell more grotesque than any medieval woodcut

#141 Post by Dougal »

HairyWill wrote:I'm having problems resizing my pup_save, it is installed on an ext3 partition.
Ok, I'll look into it.
I must say, though, that I have recently resized mine and seem to recall that it worked ok... and I don't have the unmounting problem -- just run with pfix=fsck a couple of days ago and it was ok, so maybe it is related to the partition in some way?
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind

mcewanw
Posts: 3169
Joined: Thu 16 Aug 2007, 10:48
Contact:

#142 Post by mcewanw »

Dougal wrote: I must say, though, that I have recently resized mine and seem to recall that it worked ok... and I don't have the unmounting problem -- just run with pfix=fsck a couple of days ago and it was ok
I can only imagine what I'd think, if I stumbled upon this site and didn't know a word of Puppy Linux talk; the mind boggles.

User avatar
HairyWill
Posts: 2928
Joined: Fri 26 May 2006, 23:29
Location: Southampton, UK

#143 Post by HairyWill »

Dougal,
I think this is likely to be my fault. My kernel boot line included
PUPSAVE=ext2,hda3,/pup214_save.3fs
considering both the partition and the pup_save are (I think) formatted ext3, puppy is justified in being confused.
Will
contribute: [url=http://www.puppylinux.org]community website[/url], [url=http://tinyurl.com/6c3nm6]screenshots[/url], [url=http://tinyurl.com/6j2gbz]puplets[/url], [url=http://tinyurl.com/57gykn]wiki[/url], [url=http://tinyurl.com/5dgr83]rss[/url]

User avatar
jcoder24
Posts: 604
Joined: Fri 06 May 2005, 12:33
Location: Barbados

Re: re: grubby little details

#144 Post by jcoder24 »

Dougal wrote:so I need to find a sure way of knowing where to squeeze it in.
How about looking for lines that begin with "title" --see copy of my menu.lst below.

Code: Select all

# Start GRUB global section
timeout 15
color light-gray/blue black/light-gray
# End GRUB global section

title Puppy Linux - NOP
rootnoverify (hd0,1)
kernel /puppynop/vmlinuz pmedia=idehd psubdir=puppynop
initrd /puppynop/initrd.gz

title Puppy Linux - VMPlayer
rootnoverify (hd0,1)
kernel /vmlinuz root=/dev/ram0 pmedia=idehd
initrd /initrd.gz

title Puppy Linux - Dingo
rootnoverify (hd0,1)
kernel /puppy4/vmlinuz pmedia=idehd psubdir=puppy4
initrd /puppy4/initrd.gz

title GeeXbox
rootnoverify (hd0,1)
kernel /GEEXBOX/boot/vmlinuz root=/dev/ram0 rw init=linuxrc splash=silent vga=789 video=vesafb:ywrap,mtrr
initrd /GEEXBOX/boot/initrd.gz

User avatar
Dougal
Posts: 2502
Joined: Wed 19 Oct 2005, 13:06
Location: Hell more grotesque than any medieval woodcut

Re: re: grubby little details

#145 Post by Dougal »

jcoder24 wrote:
Dougal wrote:so I need to find a sure way of knowing where to squeeze it in.
How about looking for lines that begin with "title"
I actually thought about that last night (I always end up thinking of these things when I can't fall asleep...).
The problem is that I'd rather add it after the last entry (and that it might mess up the comments a bit...).
It's good if people can show me examples of differently formatted grub menus, so I can sure I don't end messing up some funny special option included by some distros...
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind

User avatar
prehistoric
Posts: 1744
Joined: Tue 23 Oct 2007, 17:34

grub menu.lst

#146 Post by prehistoric »

@Dougal

The suggestion about title lines is about as good as anything I can come up with. This is close to the point where you should stop programming anyway. I can go through a collection of installations and produce a syntactic description of the Grub menu.lst files generated by different systems, yet still fail to deal with special cases.

SuSE linux had, at least at one time, a hash code to tell it if anyone manually edited a configuration file. No matter what you do you would change that code, causing the configuration scripts to abandon that file to manual control.

The most harmless change I can think of is to simply append the Puppy 214R entry to the menu.lst file as a comment. People, even newbies, do a good job of separating code and grouping comments. Removing comment symbols and positioning the entry where desired ought to be within the ability of any Puppy user who would think of doing a hard disk installation of any kind.

Making the operation fully automatic in all cases is the kind of thing that leads to big, powerful installers, like the PCLinuxOS installer. It also pushes you toward the "system du jour" world of endless versioning and updates. When this works it is convenient for the user; when it fails it is incomprehensible to the vast majority. On big Linux systems the feeling of fighting the installer is one of my pet peeves.

I'll try to collect some other examples when I test the Dec 22 pet above. Right now I'm involved in some seasonal activities.

(As I understand it, the tribal elders have assured us that the Sun has responded to entreaties to head north, causing everyone to go out and take part in wretched excess. This is traditional, and more economical than earlier methods of sacrificing virgins. If I have misunderstood and offended people all I can say in defense is that I'm a stranger in this millennium.)

prehistoric

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

Host misidentification problem from 2.17

#147 Post by rerwin »

Thank you, Dougal and Raffy, for an ideal Puppy -- the stability of pre-2.17 and the improvements since 2.14! I hope to replace 2.17 with it on my day-to-day PC. However, it does retain a problem I identified in 3.00beta, carried over from 2.17. It was about dhcpcd's impact on my old router.

Puppy has always identified itself with the ethernet interface's MAC address, until 2.17. The dhcpcd version used in 2.17 defaults to something for a new standard (ipv6) that impacts some older routers (like mine) that track connections by MAC address. I discussed this at length on the forum, but tempestuous netted it out here: http://www.murga-linux.com/puppy/viewto ... 983#143233

Since Puppy has always used the MAC address, by accepting the default, and has never given an option to use the ipv6 standard, I request that the net-setup script continue the practice by adding a parameter recommended by the author (but undocumented) to retain use of the MAC address. The author has violated the rule specified in the man page, by changing the default:
-I <ClientID>
Specifies the client identifier string. dhcpcd uses the default client identifier (MAC address of the network interface) if it is not specified.
Use of the MAC address can be indicated by inserting a null client-ID parameter into line 840 of net-setup.sh, resulting in:

Code: Select all

if dhcpcd -d -I '' "$INTERFACE"
EDIT 12/29: While the above change alone appears to correct the problem, for consistency, two other calls to dhcpcd should also have the "-I ''" inserted. Line 861 in net-setup.sh should be:

Code: Select all

dhcpcd -I '' ${INTERFACE}
And for rebooting, line 221 of /etc/rc.d/rc.network should be:

Code: Select all

     dhcpcd -I '' ${INTERFACE}
/EDIT

My original posts started here: http://www.murga-linux.com/puppy/viewto ... 910#141355.
However, in 3.00 beta2 I found another problem with dhcpcd in that it mistakenly showed my connection as active even though it actually had failed.
http://www.murga-linux.com/puppy/viewto ... 910#143292
So the version in 2.17 & 2.14R is not reliable; I think Barry backed it out of Puppy 3 and replaced it with a later version in Puppy 4, that uses the MAC address. Ideally, the latest version (in 4.00 alphas) would be compiled for the 2.14 kernel and used in 2.14R. But that may go beyond what is reasonable to do.

An alternate solution is to just revert to the 2.14-2.16 version of dhcpcd (1.3.22), if there are no problems with it. I substituted that version in my 2.17.1 system and right now in 2.14R. I do not know what problems others had with it.

Richard
Last edited by rerwin on Sun 30 Dec 2007, 00:54, edited 1 time in total.

mcewanw
Posts: 3169
Joined: Thu 16 Aug 2007, 10:48
Contact:

Re: Host misidentification problem from 2.17

#148 Post by mcewanw »

rerwin wrote:Thank you, Dougal and Raffy, for an ideal Puppy -- the stability of pre-2.17 and the improvements since 2.14!
I'm going to try this puppy next, it sounds very attractive to me. I've been enjoying 2.17 but I had a feeling a slightly earlier distribution would suit my old laptop even better. Looking forward to giving it a spin.

User avatar
prehistoric
Posts: 1744
Joined: Tue 23 Oct 2007, 17:34

example menu.lst generated by Mepis

#149 Post by prehistoric »

@Dougal

Here's a brand new menu.lst from an installation of SimplyMepis 7.0 for your collection of examples from other Linux systems.

Code: Select all


timeout 15
color cyan/blue white/blue
foreground ffffff
background 0639a1

gfxmenu /boot/grub/message

title MEPIS at sda1, newest kernel
root (hd0,0)
kernel /boot/vmlinuz root=/dev/sda1 nomce quiet splash vga=791 
boot

title MEPIS at sda1, previous kernel (if any)
root (hd0,0)
kernel /boot/vmlinuz.old root=/dev/sda1 nomce quiet splash vga=791 
boot

title MEPIS at sda1, kernel 2.6.22-1-mepis-smp
root (hd0,0)
kernel /boot/vmlinuz-2.6.22-1-mepis-smp root=/dev/sda1 nomce quiet splash vga=791 
boot

title TinyFlux linux 1.0
root (hd0,2)
kernel /boot/vmlinuz BOOT_IMAGE=linux root=/dev/hda3 acpi=on resume=/dev/hda2 splash=silent vga=788
initrd (hd0,2)/boot/initrd.img

title Puppy Linux Fire Hydrant, frugal (in hda1)
rootnoverify (hd0,0)
kernel /puppy301FH/vmlinuz pmedia=idehd psubdir=puppy301FH
initrd /puppy301FH/initrd.gz

title MEMTEST
kernel /boot/memtest86+.bin

The Puppy entry is one I added manually. The entry for Tiny Flux 1.00 was automatically extracted from the previous menu.lst and included when Mepis 7.0 installed Grub.

For completeness, here's the menu.lst generated by Tiny Flux, (PCLinuxOS installer). The Puppy entry here was also manually edited. The Windows entry refers to a system no longer present on hda1.

Code: Select all

timeout 10
color black/cyan yellow/cyan
gfxmenu (hd0,2)/usr/share/gfxboot/themes/pcfluxboxos/boot/message
default 0

title TinyFlux linux 1.0
kernel (hd0,2)/boot/vmlinuz BOOT_IMAGE=linux root=/dev/hda3  acpi=on resume=/dev/hda2 splash=silent vga=788
initrd (hd0,2)/boot/initrd.img

title TinyFlux linux-nonfb
kernel (hd0,2)/boot/vmlinuz BOOT_IMAGE=linux-nonfb root=/dev/hda3  acpi=on resume=/dev/hda2
initrd (hd0,2)/boot/initrd.img

title TinyFlux failsafe
kernel (hd0,2)/boot/vmlinuz BOOT_IMAGE=failsafe root=/dev/hda3  failsafe acpi=on resume=/dev/hda2
initrd (hd0,2)/boot/initrd.img

title Puppy Linux Fire Hydrant, frugal (in hda1)
rootnoverify (hd0,0)
kernel /puppy301FH/vmlinuz pmedia=idehd psubdir=puppy301FH
initrd /puppy301FH/initrd.gz

title windows
root (hd0,0)
makeactive
chainloader +1
I'll pick up a few more examples over the next few days.

prehistoric

User avatar
Dougal
Posts: 2502
Joined: Wed 19 Oct 2005, 13:06
Location: Hell more grotesque than any medieval woodcut

Re: Host misidentification problem from 2.17

#150 Post by Dougal »

rerwin wrote:Use of the MAC address can be indicated by inserting a null client-ID parameter into line 840 of net-setup.sh, resulting in:

Code: Select all

if dhcpcd -d -I '' "$INTERFACE"
I've implemented that, I'll have to ask Tempestuous about updating/reverting the dhcp version.
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind

User avatar
Dougal
Posts: 2502
Joined: Wed 19 Oct 2005, 13:06
Location: Hell more grotesque than any medieval woodcut

Re: example menu.lst generated by Mepis

#151 Post by Dougal »

prehistoric wrote:@Dougal

Here's a brand new menu.lst from an installation of SimplyMepis 7.0 for your collection of examples from other Linux systems.
Thanks, I guess I'll just go for the "title" lines (seems like most installers don't use comments anyway, so I won't be messing anything up.
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

Re: Host Misidentification Problem

#152 Post by rerwin »

Thanks, Dougal, for accepting my MAC address client-ID fix. I have discovered two other instances of dhcpcd calls that should be changed, too. I have edited my original post with the following:
While the above change alone appears to correct the problem, for consistency, two other calls to dhcpcd should also have the "-I ''" inserted. Line 861 in net-setup.sh should be:

Code: Select all

dhcpcd -I '' ${INTERFACE}
And for rebooting, line 221 of /etc/rc.d/rc.network should be:

Code: Select all

     dhcpcd -I '' ${INTERFACE}
Be aware that this fix is appropriate for only dhcpcd 3.14. If the older dhcpcd is substituted, the fix causes problems -- you don't want to go there.

EDIT re 3.1.8: This fix also works with dhcpcd 3.1.8, used in the 4.00 alphas, although the line numbers to change are different. I experimented with making the choice optional, and now have it working except for a user interface. Since WinXP and Ubuntu still use the MAC address-only client ID, I see no benefit in adding the new client-ID format before the mainstream OSes do. Another reason to use the fix is that the new format results in a new IP address assignment for each pup_save file/partition. Also, after setting up networking without a pup_save and then saving to a new pup_save, the reboot to start using the pup_save results in a new IP address being obtained, different from that gotten before the reboot.
Richard

User avatar
Dougal
Posts: 2502
Joined: Wed 19 Oct 2005, 13:06
Location: Hell more grotesque than any medieval woodcut

#153 Post by Dougal »

I've implemented using the "title" line in the grub menu for adding the new entry, if anyone wants to try it...
Attachments
puppyinstaller.gz
(19.57 KiB) Downloaded 405 times
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind

Post Reply