EasyOS version 2.3.2, June 22, 2020

[foxpup]

halpad
If the rollback does not help...
maybe it hangs because it cannot start X. I have this problem with an older PC and EasyOS. In that case you first choose noX (is it in the boot menu?) and on the prompt you continue with xwin.

[rufwoof]

Downloaded, unzipped and opened the .img file, copied the initrd.q/vmlinuz/q.sfs to mechanical HDD /easy/easy-0.91 folder and set grub4dos menu.lst to boot it i.e. frugal install (note I'm using find --set-root .... rather than coding in a UUID)

Code: Select all

# inside initrd.q BOOTSPEC use boot disk id from fdisk -l /dev/sda
title Easy 0.91 (frugal install in sda2)
 find --set-root /easy/easy-0.91/vmlinuz
 kernel /easy/easy-0.91/vmlinuz
 initrd /easy/easy-0.91/initrd.q

Using Easy 0.9 clicked on the initrd.q to open it up and it auto-detected the BOOTSPEC needed to be changed and auto made changes, however I had to further edit it to correct it to

Code: Select all

BOOT_DISKID='0x00055e48'
BOOT_PARTNUM=2
BOOT_FS='ext3'
BOOT_DIR='easy/easy-0.91/'
WKG_DISKID='0x00055e48'
WKG_PARTNUM=2
WKG_FS='ext3'
WKG_DIR='easy/easy-0.91/'
Q_DISTRO_VERSION=0.9.1

i.e. filesystem (FS) type and BOOT_DIR/WKG_DIR paths.

First boot I had to run xorgwizard to select my Radeon graphics. After firstrun setup I used JWM Desk Manager to set screen font to 132. Looks like Pup Control Panel code only caters for up to 120, so when displayed its compacted and much less usable (have to horizontally scroll to see the Pup Control Menu box choices).

Container creation - if you set the tick values before selecting which application is to be containered from the drop down box has all of those ticks reset back to the default settings again after selecting the application. Default choices for the ticks are also a bit 'weird'.

Set up a sakura container, set up rover home folder within that (.profile, .gtkrc-2.0 ...etc. and added rover to audio and video groups) and downloaded/extracted firefox 59.03 Quantum within the container. Also installed Oscar's 64 bit apulse and created a user.js files for that. Set rover passwd and logging in as rover and running firefox and videos/sound all working fine.

Personally I like the blue theme and prefer it over 0.9 green theme.

[rufwoof]

Downloaded and installed the audacity PET from this post by watchdog and after 2 or 3 sym linking libs its working well in Pyro 0.91

Don't know if LD_LIBRARY_PATH is required or not? none set in the base system but I added one to /root/.bashrc containing

Code: Select all

LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/lib:/usr/lib:/usr/lib/x86_64-linux-gnu export LD_LIBRARY_PATH

Used gparted to wipe a USB, creating a FAT32 formatted stick, used the installed Audacity to convert some mp4's to mp3's (and amplify one some as it was comparatively quiet compared to the others), copied the mp3's to the USB and they're playing fine when plugged into the cars USB.

[Cu Chulinux]

Ah ok ... rollback means start over. I thought maybe it was related to the repository and roll back to a previous version.

So am I right that first time Easy runs it creates the .session dir and saves some hardware settings or does it autodetect the hardware every time it boots? I haven't really looked into .session to see what's in there yet.

[upnorth]

.session dir was the only place I could find root, etc, user, dirs using only rox. I made some changes to etc and root and they persist. Other than that,

0.9.1, tried first on this older machine:

AMD Phenom(tm) II X2 B55 Processor

Actual Used RAM: 157 MB Used - (buffers + cached)
Actual Free RAM: 3541 MB Free + (buffers + cached)

Distro: Easy Pyro64 0.9.1

• Xorg Driver in use: radeon
• Loaded Modules: dbe dri2 evdev exa extmod fb glx ramdac
• X.Org version: 1.19.1

OpenGL 2D/3D Rendering:
• Direct Rendering: Yes
• Vendor: X.Org
• Renderer: Gallium 0.4 on AMD RS880 (DRM 2.50.0 / 4.14.39, LLVM 3.9.1)
• Version: 2.1 Mesa 17.0.7

Live Hd video worked fine in Seamonkey - containerised . No problems after a few hours.
Added ublock origin to seamonkey.

[rufwoof]

Ah ok ... rollback means start over. I thought maybe it was related to the repository and roll back to a previous version.

So am I right that first time Easy runs it creates the .session dir and saves some hardware settings or does it autodetect the hardware every time it boots? I haven't really looked into .session to see what's in there yet.

I mentally visualise it as q.sfs (the main 'puppy' sfs) at the bottom, and .xsession as a save folder overlaid on top of that. i.e stores all the changes in .xsession.

Create a container and it creates another layered filesystem using the same main sfs at the bottom level, but another .xession save folder on top of that.

Make a snapshot of a container (or the main system) and it coverts the associated .xsession into a sfs, so later you can in effect wipe out the current .xsession and replace it with the content of that snapshot sfs (i.e. rollback) or whatever. But where you can also rollback to a later version (roll-forward again) i.e. replace the .xsession again but with a later snapshot sfs copy of .xsession).

So you can rollback either the main system, or containers to any prior snapshot (sfs copy of .xsession), or even forward again. I think the rollback to what you are referring is the bootup choice i.e. rollback to the first snapshot of the main version that is automatically created after the first run. Or you can boot as though rolling back to the very first run, no .xsession overlay content at all initially.

And of course you can add in other sfs's into that mix as well.

Fundamentally I love Easy as I can run the browser (main security risk) in a container as a restricted userid etc. and run everything else (non internet) as a normal puppy style root and the flexibility/ease that provides. So I only need the single container (I set up a sakura/terminal container for that as I like to run firefox which I install inside the container, but being a terminal session container I can also tweak things from inside the container (I'm using capsh to restrict things further along with setting file/folder permission such that even root inside the container is pretty much useless, but in a manner where the browser still runs fine). Accordingly I'm less inclined to use the snapshots and rollbacks and instead I'm just running normal backups from the 'outside' (copying the frugal install directory for EasyOS).

[greengeek]

Adding rover to group video gets rid of the messages.

Sorry for my ignorance - could you suggest how this is done please? cheers!

[rufwoof]

Adding rover to group video gets rid of the messages.

Sorry for my ignorance - could you suggest how this is done please? cheers!

Once you've created a container for sakura, open that and it will be running as root with q.sfs as the bottom layer and its own containers .xsession folder at the top to store changes. Very much like running a normal session. As root running addgroup audio rover for instance will set user rover belonging to both netshare and audio groups (its set to be in group netshare as part of the standard build).

What I then did was copy .gtkrc-2.0 ... and whatever over from /root to /home/rover within the container and then chown'd all files under that chown -R rover:netshare /home/rover along with chmod'ing /root /mnt so rover can't access /root or /mnt and also set a password for rover using passwd rover ... and entering a password twice.

So conceptually user rover can be run as though a normal user, but its running inside a container with its own q.sfs and .xsession file/folder stack.

I then created a .bashrc in the containers /root folder where my current content for that looks like

Code: Select all

#############################################################
#
# NOTES
#
#CAP_NET_ADMIN - Perform various network-related operations - interface configuration - administration of IP firewall, masquerading and accounting - modify routing tables - bind to any address for transparent proxying - set type-of-service (TOS) - clear driver statistics - set promiscuous mode - enabling multicasting - use setsockopt() for privileged socket operations
#CAP_NET_BIND_SERVICE - Bind a socket to Internet domain privileged ports (less than 1024)
#CAP_SYS_ADMIN - Very powerful capability, includes: - Running quota control, mount, swap management, set hostname, ... - Perform VM86_REQUEST_IRQ vm86 command - Perform IPC_SET and IPC_RMID operations on arbitrary System V IPC objects - Perform operations on trusted.* and security.* extended attributes - Use lookup_dcookie
#CAP_SETPCAP - Allow the process to add any capability from the calling thread's bounding set to its inheritable set, and drop capabilities from the bounding set (using prctl()) and make changes to the securebits flags.
#CAP_SYS_BOOT Use reboot() and kexec_load()
#CAP_SYS_CHROOT Use chroot()
#CAP_SYS_MODULE -     Load and unload kernel modules
#CAP_SYS_RESOURCE - Another capability with many consequences, including - Use reserved space on ext2 file systems - Make ioctl() calls controlling ext3 journaling - Override disk quota limits - Increase resource limits - Override RLIMIT_NPROC resource limits
#
#############################################################

# Option here to access the container as a normal root type session for admin purposes

echo "Press the 'r' key to enter root userid mode."
echo "Any other key or wait 2 seconds for rover userid mode"
read -t 2 -n 1 R
if [ "$R" = "r" ]; then # ensure permissions of files are 'normal' and drop to standard shell
  DISPLAY=:0 export DISPLAY
#  chmod 755 /usr/sbin
#  chmod 755 /sbin
#  chmod 755 /bin/busybox
#  chmod 755 /bin/chmod
#  echo "if you're seeing errors from chmod permissions then you haven't deleted"
#  echo "something like /mnt/wkg/containers/sakura/.session/bin/chmod"
#  echo "so that prior q.sfs chmod permissions show through to the top layer"
  /bin/sh
  exit
fi

#########################################################
#
# Or we drop into a rover userid shell, where root is extremely limited bordering on 
# ineffectual, so even if a browser or other break out occurs elevating privileges is
# difficult and even if root level authority in the container is achieved that is pretty
# much useless 
#
#########################################################
# Note with the permissions we set, tail and tr functions are unavailable (being in busybox)
# and /etc/profile sets the DEFAULTBROWSER ... etc. type environment variables using those
# functions so in this container they are all empty i.e. runs
#DEFAULTBROWSER="`cat /usr/local/bin/defaultbrowser | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTDRAW="`cat /usr/local/bin/defaultdraw  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTHTMLEDITOR="`cat /usr/local/bin/defaulthtmleditor  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTMEDIAPLAYER="`cat /usr/local/bin/defaultmediaplayer  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTPAINT="`cat /usr/local/bin/defaultpaint  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTSPREADSHEET="`cat /usr/local/bin/defaultspreadsheet  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTTEXTEDITOR="`cat /usr/local/bin/defaulttexteditor  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTWORDPROCESSOR="`cat /usr/local/bin/defaultwordprocessor  | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTIMAGEVIEWER="`cat /usr/local/bin/defaultimageviewer | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#DEFAULTIMAGEEDITOR="`cat /usr/local/bin/defaultimageeditor | tail -n 1 | tr -s " " | cut -f 2 -d " "`"
#export DEFAULTBROWSER DEFAULTDRAW DEFAULTHTMLEDITOR DEFAULTMEDIAPLAYER DEFAULTPAINT DEFAULTSPREADSHEET DEFAULTTEXTEDITOR DEFAULTWORDPROCESSOR DEFAULTIMAGEVIEWER DEFAULTIMAGEEDITOR
. /etc/profile >/dev/null 2>&1 # avoid showing all the tr and tail errors
#########################################################

#v1.0.5 need to override TERM setting in /etc/profile...
#export TERM=xterm
# ...v2.13 removed.

#export HISTFILESIZE=2000
#export HISTCONTROL=ignoredups
#...v2.13 removed.

#Number   SIG   Meaning
#0   0   On exit from shell
#1   SIGHUP   Clean tidyup
#2   SIGINt   Interrupt
#3   SIGQUIT   Quit
#6   SIGABRT   Abort
#15   SIGTERM   Terminate

trap finish 0 1 2 3 6 15
finish()
{
  exit
}

# Haven't got this working yet as permission denied to create
# folders under /sys/fs/cgroup !!! 
#mkdir /sys/fs/cgroup/memory
# limit container memory to 1GB
#echo "100000000" > /sys/fs/cgroup/memory/memory.limit_in_bytes
# Limit CPU share to 50% (1024 being 100%)
#echo 512 > /sys/fs/cgroup/memory/cpu.shares
# set the current PID to adopt that group
#echo $$ > /sys/fs/cgroup/memory/tasks 


# May already be set, so redirect so stderr not showing
#chmod 700 /root >/dev/null 2>&1
#chmod 000 /usr/sbin >/dev/null 2>&1
#chmod 000 /sbin >/dev/null 2>&1
#chmod 000 /bin/busybox >/dev/null 2>&1
#chmod 000 /bin/chattr.e2fsprogs >/dev/null 2>&1
#chmod 000 /bin/dd >/dev/null 2>&1
#chmod 000 /bin/kmod >/dev/null 2>&1
#chmod 000 /bin/login >/dev/null 2>&1
#chmod 000 /bin/mount >/dev/null 2>&1
#chmod 000 /bin/mount-FULL >/dev/null 2>&1
#chmod 000 /bin/pupkill >/dev/null 2>&1
#chmod 000 /bin/umount >/dev/null 2>&1
#chmod 000 /bin/umount-FULL >/dev/null 2>&1
#chmod 000 /bin/chmod >/dev/null 2>&1
chmod o-wrx /mnt /root /sbin /bin/busybox /bin/kmod /usr/sbin /bin/chattr.e2fsprogs /bin/login \
 /bin/mount /bin/mount-FULL /bin/pupkill /bin/umount /bin/umount-FULL /bin/chmod


HOME=/home/rover export HOME
DISPLAY=:0 export DISPLAY
XDG_DATA_HOME=/home/rover/.local/share
GTK2_RC_FILES=/home/rover/.gtkrc-2.0
USER=rover
PATH=/bin:/usr/bin:/sbin:/usr/sbin:/usr/local/bin
XFINANSDIR=/home/rover/.xfinans
XDG_CONFIG_HOME=/home/rover/.config
XDG_CACHE_HOME=/home/rover/.cache
LOGNAME=rover
cd /home/rover

capsh --drop=cap_chown,cap_sys_admin,cap_sys_chroot,cap_net_admin,

cap_net_bind_service,cap_sys_boot,cap_sys_module,

cap_sys_resource,cap_setpcap,cap_setgid,cap_setuid,

cap_sys_rawio,cap_mknod,cap_net_raw,cap_audit_control,

cap_mac_override,cap_mac_admin,cap_syslog,cap_audit_read,

cap_audit_write,cap_fsetid --keep=1 --user=rover --uid=1003 -- -c /bin/sh -- 

exit
#############################################################

NOTE : I've separated the capsh single line into multiple lines in the above otherwise the forums thread width expands out to be way too wide a display. i.e. all of those capsh values starting from capsh --drop= are on a single line.

So when the container starts (and is running as root) it rolls through that and either runs /bin/sh if r is pressed within 2 seconds or otherwise falls through to run /bin/sh as rover that has been capsh'd quite heavily. So if I want to install firefox I'd use that r option and install (extract) firefox under /usr/libs ... or otherwise don't use the r option and just run firefox as rover (run /usr/lib/firefox/firefox).

To get sound working you have to install Oscar's apulse and edit/create a user.js file under /home/rover/.mozilla/... sub folder (as per how other threads describe to get apulse and firefox sound working).

Very much a case of rover (or it could be spot or fido) can be made to actually work in a 'puppy' now.

In my prior 0.9 version, as part of that .bashrc I severely crippled the container such as setting /sbin to no permissions even for root along with other files/folders, but that's only appropriate once you've set everything up as afterwards its pretty fixed/difficult to change. That way even if a privilege elevation from rover to root occurred inside the container, then that's still pretty much useless. You could reset the permissions from the main session however i.e. something like chmod'ing /mnt/sda2/easy/easy-0.91/containers/sakura/.session/sbin ... or simply just delete sbin in that folder so the q.sfs version shines through as the top level.

EDIT: Nearly forgot. I also create a .profile in the containers /home/rover folder containing

Code: Select all

DISPLAY=:0 export DISPLAY
HOME=/home/rover export HOME
XDG_DATA_HOME=/home/rover/.local/share export XDG_DATA_HOME
GTK2_RC_FILES=/home/rover/.gtkrc-2.0 export GTK2_RC_FILES
XFINANSDIR=/home/rover/.xfinans export XFINANSDIR
XDG_CONFIG_HOME=/home/rover/.config export XDG_CONFIG_HOME
XDG_CACHE_HOME=/home/rover/.cache export XDG_CACHE_HOME

to as to set up its environment.

That could all be created and made into a sfs I guess, so a basic rover command prompt could be more easily shared around/installed by others. Also the capsh command could be set to run firefox by default rather than /bin/sh i.e. the second parameter from the end of that long command line.

[Cu Chulinux]

rufwoof, I have been reading your posts with great interest. Very interesting and informative. I haven't quite got the inner details of Easy figured out yet (just too busy with my job right now) but I agree that Easy has amazing potential. I'm waiting for the "Aha" moment when it all makes sense to me. It's close.

I need to play around with containers myself when I have more time.

[rufwoof]

A squashed file system (SFS) can be likened to a zip file, a single file that holds compressed versions of multiple files. Like a zip file you can pull out any one or combination of files, but only in a read only like manner. To add or remove things you have to rebuild the zip/sfs. In the case of SFS the kernel is extended so that it knows how to handle the sfs content as though it were a file system similar to any other filesystem.

To make it more dynamic you overlay that read only system with a folder (keeping things simple). So if a new file is added it is recorded/added to that folder. And where what the end user 'sees' is that folder at the top level (takes precedence) over the lower read only layer. For changed files, the later version in the folder is seen rather than the one in the read only layer. For deleted files a special whiteout (.wh filename suffix) is used, so if there is a file called /abc in the read only layer but a abc.wh file exists in the folder layer then the system treats it as though that /abc file doesn't exist.

Initially the changes folder layer is empty, but as files are changed/added/deleted so it grows in size. Ultimately I guess the folder layer would contain just as much as the sfs, but it would take time for that to occur and you could create a new bottom read only sfs from the combined current sfs and folder view so as to have a new bottom read only sfs and a empty folder layer again.

A EasyOS container is a neat trick where it uses the same sfs as the bottom layer, but a different folder as its top layer (changes), so initially it takes up near no extra space to create. Barry has built/structured it so you could have a container for each separate running process in a similar vein to how others run things in separate containers for security purposes. I'm however more inclined to just have a single container i.e. use the main system that runs as root for local non internet activities (word processing, image editing/whatever), but use the container for internet based activity and where that container is locked down pretty tightly. As I use online email, that pretty much just leaves the browser being in a container for me. Everything else being run as normal (as root). Once set up I have a rox bookmark that links into that container so I can move files between the main system and into/out of the container using rox. In effect everything running as root, except the greatest security risk i.e. the browser that runs in a restricted environment and runs as a restricted user id (rover in my case).

Puppy security (running as root, outdated software and browser etc.) were issues for me before. EasyOS goes a considerable way to eradicating such issues. Seems such a easy concept that you wonder why it wasn't common/used before. But that's (Barry's) genius for you, making things so easy it looks as simple as abc or e=mc squared. Barry's brain however does see light differently to mine, more often what he sees as a nice colour combination looks awful to me and vice-versa. Easy Pyro 0.91 blue colour theme that Barry is unsure about for instance looks vastly better to me than the black text on dark green colour theme that Barry liked in 0.9 but that looked horrible to me Clearly our minds are on different wavelengths.

[rufwoof]

crond (crontab daemon) is by default disabled in Puppy, but can be started and Puppy also has Pshedule. crontab is both a actual executable and has a busybox version in EasyOS.

Is is possible to set permissions on individual busybox commands such as crontab?

One risk factor/means of entry is to pirate/inject javascript (cross scripting) that activates core dumping and cd to /etc/cron.d ... and crash, so that a core dump is created that contains code the hacker wishes to execute as root. If crond is running that will interpret many of the cron'd commands in the core dump as "can't do" type (invalid) actions but where one or more of the commands are valid and that were injected by the hacker i.e. who then has root level access to send out requests and retrieve further commands/actions to run as root.

Turning off crond and ensuring it can't be run/activated is therefore a desirable protection within a browser type container. But short of totally disabling busybox within the container (firefox still apparently seems to run OK) it would be nicer if it could be more refined (selective busybox commands having permissions set to 000).

[rameshiyer]

This is my BOOT_SPEC.
BOOT_DISKID='c1a29569-e75d-4d88-a55f-ae46693767e3'
BOOT_PARTNUM=1
BOOT_FS='ext4'
BOOT_DIR='Easy'
WKG_DISKID='c1a29569-e75d-4d88-a55f-ae46693767e3'
WKG_PARTNUM=1
WKG_FS='ext4'
WKG_DIR='Easy'
Q_DISTRO_VERSION=0.9.1

My Hard disk details ;
# fdisk -l /dev/sda
Disk /dev/sda: 465.8 GiB, 500107862016 bytes, 976773168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0xa2d29de5

Device Boot Start End Sectors Size Id Type
/dev/sda1 * 2048 206847 204800 100M 7 HPFS/NTFS/exFAT
/dev/sda2 206848 539473919 539267072 257.1G 7 HPFS/NTFS/exFAT
/dev/sda3 539473921 854015399 314541479 150G f W95 Ext'd (LBA)
/dev/sda5 539473984 612879749 73405766 35G 83 Linux
/dev/sda6 612879813 665299844 52420032 25G 83 Linux
/dev/sda7 665299908 717719939 52420032 25G 83 Linux
/dev/sda8 717720003 770140034 52420032 25G 83 Linux
/dev/sda9 770140098 854015399 83875302 40G 7 HPFS/NTFS/exFAT

Grub4Dos menu List

# menu.lst produced by grub4dosconfig-v1.9.2
color white/blue black/cyan white/black cyan/black
#splashimage=/splash.xpm
timeout 10
default 0

# Full installed Linux

title Quirky Xerus64 #QUIRKY 8.3 #QUIRKY (sda6/boot)
uuid c7d57fa9-e3b5-4413-ad85-44f5f0b5d01a
kernel /boot/vmlinuz root=/dev/sda6 ro

title Xenialpup 7.5 (sda7/boot)
uuid f5234c5e-dbe3-4251-a281-56d7bd51a5e5
kernel /boot/vmlinuz root=/dev/sda7 ro

title Easy Linux (sda8)
uuid c1a29569-e75d-4d88-a55f-ae46693767e3
kernel /Easy/vmlinuz
initrd /Easy/initrd.q

# Windows
# this entry searches Windows on the HDD and boot it up
title Windows\nBoot up Windows if installed
errorcheck off
find --set-root --ignore-floppies --ignore-cd /bootmgr
chainloader /bootmgr
find --set-root --ignore-floppies --ignore-cd /ntldr
chainloader /ntldr
find --set-root --ignore-floppies --ignore-cd /io.sys
chainloader /io.sys
errorcheck on

# Advanced Menu
title Advanced menu
configfile /menu-advanced.lst
commandline

Frugal HDD Installation. ( Created directory Easy in sda8 ( put the three file). Still not able to boot the Easy OS. Showing error while booting that finding sda etc.. Boot drive not found.
Kindly pin point where is the problem in my setup. Also note that I was able to put up through flash disk. I have updated initrd.q through Easy Os booted through Pen Drive.

[rufwoof]

For one your Boot and working dir's need trailing slash

BOOT_DIR='Easy/'

WKG_DIR='Easy/'

The PARTNUM for both should also be 8 ?? And perhaps that partition isn't a ext4 format filesystem?? (Not ext3 or ext2)

Perhaps also try changing the menu.lst entry for Easy to

title Easy Linux (sda8)
#uuid c1a29569-e75d-4d88-a55f-ae46693767e3
find --set-root /Easy/vmlinuz
kernel /Easy/vmlinuz
initrd /Easy/initrd.q

[Cu Chulinux]

Yes, partnum should be 8. The error is from looking for the initrd.q (or maybe q,sfs) file on sda1 which it is told is an ext4 partition and can't find.

(edit) it would have to be q.sfs since initrd.q would be loaded by the kernel based on grub4dos parameters. Bootspecs is unavailable until after initrd.q is loaded (Doh!).

[rameshiyer]

Dear Sir (Rufwoof)

Thank you very much pin pointing the issue. I am very happy to inform you that now the my problem solved. As suggested by you, I have added trailing slash (BOOT_DIR='Easy/' and WKG_DIR='Easy/' ), it solved my issue. Finally I was able to run the frugal install from hard disk. Now I am posting from Easy OS 0.9.1.

No static IP setting in the LAN Network. Somehow I have managed with Frisbee.

Thank you once again for the support.

[scsijon]

Sorry Barry, a curly one for you.

Decided to install 0.9.1 to a spare desktop to try to build a package that I ocassionally play as a stress relief (ToME).
Installed 0.9.1 to a memory stick with dd ok.
Blew a spare hard drive away and put it in one of my dev desktops.
0.9.1 installed to the hard drive sda1 ok from the memory stick.
shutdown the desktop and removed the memory stick.
I booted the hard drive up and it added sda2 with all spare space.
I wanted to add a swap drive and put the memory stick back in and rebooted the desktop.
I told the pc (via f12) to boot the memory stick and not the hard drive.
The memory stick light came up once, however in starting it autopicked to start from sda/sda2, not sdb1/sdb2 as I had expected it to do.
Shutdown the desktop and tried starting it again, and got the same result.
I got out an earlier pyro (0.8.1) memory stick and tried again (with F12) and it booted correctly.
Blew the hard drive clean and tried it all with 0.8.1 and it worked correctly, both the install and restart from the memory stick.
cleaned the hard drive out again.
tried again with 0.9.1 and got the same problem.
It seems that I can install it once, but restarting a second time from the memory stick it's automatically picking the hard drive to load and start from.

Any ideas?

[BarryK]

Hi guys,
I will read the above posts carefully soon.

Right now, getting "Easy Beaver" to work properly.

This is the same Easy OS, but built with Ubuntu 18.04 Bionic Beaver DEBs. This is great for those who want the big repos in the PPM.

Have a desktop, but there are things to fix. And to compile, the kernel for instance needs to be at least 4.15.

My monthly data limitation is now solved:

http://bkhome.org/news/201805/telstra-6 ... -plan.html

[BarryK]

Easy Beaver 0.9.2 is out!

Blog post here:

http://bkhome.org/news/201805/easy-beav ... eased.html

I ended up using kernel 4.16.8.

If you want to try it, you're welcome. Especially if you fix any bugs

[Billtoo]

I installed to a 16gb SDHC card:

System: Host: EASYPC19814 Kernel: 4.16.8 x86_64 bits: 64 Desktop: JWM git-1685 Distro: Quirky 0.9.2 beaver64
Machine: Device: desktop System: Compaq-Presario product: AU194AA-A2L CQ5123F serial: MXX9300M0F
Mobo: MSI model: Boston v: 1.0 serial: N/A BIOS: Phoenix v: 5.24 date: 06/19/2009
CPU: Dual core Pentium E5200 (-MCP-) speed/max: 1523/2500 MHz
Graphics: Card: Intel 82G33/G31 Express Integrated Graphics Controller
Display Server: X.Org 1.19.6 driver: intel Resolution: 1920x1080@60.00hz
OpenGL: renderer: Mesa DRI Intel G33 version: 1.4 Mesa 18.0.0-rc5
Network: Card: Realtek RTL8101/2/6E PCI Express Fast/Gigabit Ethernet controller driver: r8169
Drives: HDD Total Size: 516.0GB (1.9% used)
Info: Processes: 96 Uptime: 1:06 Memory: 207.8/3943.0MB Client: Shell (bash) inxi: 2.3.56

I updated PPM and installed some applications.
I'm still running internet applications as spot, will try running
Chromium in a container later on.

That's it so far,
Thanks.

[rameshiyer]

For updating 0.9.1. to 0.9.2 (Beaver Easy) which files to be replaced in the Frugal Installation folder.