Fatdog64-720 and 721 Final [11 Jan 2018]

[FanDog]

Awesome! Couldn't wait for the new release <3 Will be putting it to the test briefly.

will be visible when you run gslapt and you can remove it from gslapt. Any other behaviour is a bug. If you find this, can you please report

One possible bug was that, while uninstalling seemed to work, the menu entry was not removed. Although this could be related to the following 'bug' (which I came here to report and saw 720 under the xmas tree \o/ )

There have been a few files giving input/output error (cant delete them, nor even stat them), one of them related to a package I tried to install. Now, even gslapt wont load (it complains it cannot find said package). This is 710, just to be clear. (sorry to post here, but the goal is to request a new feature, not fix this)

It could be hardware error but I suspect it was some interaction with the Fatdog merging process. Either way the problem is, rebooting with savefile=none and doing a fsck says the FS is clean, I was even tempted to try to umount it and manually recover a superblock backup, as per the next link. (which is pretty much the problem Im having) but it was already getting too dangerous for my skillset (running it on all those overlayed, cryptoed, ram mounts?! ) Now that 720 is out, I'll simply migrate instead

https://ubuntuforums.org/archive/index. ... 45536.html

okay, so this whole post was basically a request to add such capability to the savefile utility, AND to thank you all for the work put into this great distro! My favorite (pre-compiled ) distro. My2018 couldn't have began better!

Cheers.

[step]

Having enabled BC-wl wlan0 is in wired pane on wpagui with no way to connect. Used Wifi-Scanner-2 (Ad) script only and voila! problem solved...

Great that you can connect using your script! The wireless interface shouldn't show up in the "wired" pane to begin with, so let's dig deeper as to why it happens.

Could you please go back to that system and:
1) Confirm that wlan0 still shows up in the wrong pane (wired)
2) Do _not_ connect through Wifi-Scanner-2
3) Open terminal, run the following commands and post results. Thank you.

Code: Select all

ifconfig wlan0
readlink -f /sys/class/net/wlan0/device/driver/module
find /etc/wpa_gui -print0 | xargs -0 grep '"";
find -H /sys/class/net/wlan0 \! -type d -print0 | xargs -0 grep ""

In the screenshot you posted I can see that wlan0 did actually connect to a DHCP server that allocated an IP address for wlan0. Whether that makes any sense in your environment I can't tell. Was the screenshot taken _before_ or _after_ connecting through your script?

What's the "Ad" in "Wifi-Scanner-2 (Ad)" stand for?
What's "BC-wl" in your post?

Thank you.

[stemsee]

Yes it still shows in wired selector.

The ip is there after connecting with my script.

BC-wl is the fatdog service to load the module.

I could not get the last two commands to work in xterm.

the first two returned the following.

Code: Select all

ifconfig wlan0
wlan0     Link encap:Ethernet  HWaddr 38:b1:db:d3:cc:33  
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:7191 errors:0 dropped:195 overruns:0 frame:21273
          TX packets:4892 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:7838748 (7.4 MiB)  TX bytes:608994 (594.7 KiB)
          Interrupt:18 

/sys/module/wl

Ad - advertisement!

[stemsee]

The wlan0 interface is now in the wifi pane! I think that when it is discovered in X it goes to wired pane, but during boot with a savefile it gets put into the wireless pane! For a while there was somewhere code flashing in the wpa_gui pane then it settled down and I was able to connect.

[stemsee]

I found also that geany 1.31 has a bug which prevents executing in xterm the script which is being edited. I compiled and installed geany 1.32 from source and the problem has gone.

[jake29]

I'm assuming this is related to the moving of Spot. Can anyone suggest how to deal with apps - in my case latest versions of Chrome and Opera - that now refuse to run as Root.

Code: Select all

# opera
[0105/000046.293869:ERROR:zygote_host_impl_linux.cc(88)] Running as root without --no-sandbox is not supported. See https://crbug.com/638180.

Fortunately with latest stable build of Opera (50.0.2762.45) - it will launch with --no-sandbox. Chrome refuses to.

What is the correct way to install these apps now?

[step]

I could not get the last two commands to work in xterm.

I figured you couldn't. Would you please post the error messages from those two commands? Thank you, also for posting the other stuff.

[step]

@jake29, I run opera as an sfs, pretty much as we discussed in a previous thread. However, I run it as spot, from a command line run: run-as-spot opera &
Of course, this will start a fresh profile under spot's home. You can try to port root's profile to spot's home this way: (as root)

Code: Select all

cp -a ~/.cache/opera ~spot/.cache/
cp -a ~/.config/opera ~spot/.config/
chown -R spot:spot ~spot/.cache/opera ~spot/.config/opera
run-as-spot opera &

Then change the downloads folder path in opera settings to /home/spot/Downloads.
Spot needs to own all opera cache and profile files in spot's home at all times. It doesn't need to own symbolic links.
Pre-downloaded files in spot's Downloads folder should already be under spot's ownership.
There is a slight inconvenience in running the browser as spot because the browser can't access folders under /root (including ~/Downloads). However, it should be more secure.

[stemsee]

Zarfy works much better than before in fd-720. It actually auto detects hdmi and activates it. When first plugged in it extends the display to 2nd display resolution (4k tv). If then restartX I get mirrored displays both at primary display res (1080). For me this is great, as i don't actually need to start up zarfy as previously

I also discovered that the init scripts and boot process has changed significantly (system-init) > (busybox-init (binary)). . . making it difficult to boot the main.sfs with any other initrd/init script. Also there are significant changes in the use of permissions or ACLs on root directories, it seems. Previous 710 would break with the latest upgrades using gslapt.

720 is working great now, and I wonder if all the technical changes are listed somewhere?

[dr. Dan]

How small can a Fatdog64 installation partition be and still function? Either with save folder there or elsewhere?
What is ideal as a minimum?
Thanks,
Dan

[snayak]

Dear All,

I am using Pidgin 2.11 on fatdog720.
I am unable to logging into pidgin using my gmail account.
It throws invalid ssl certificate error. https://developer.pidgin.im/ticket/17118

Did you get success using your gmail id?

Sincerely,
Srinivas Nayak

[snayak]

I found pidgin 2.10.11 in fatdog 720 repo. It worked!

[SFR]

@Snayak: they say it's fixed in v2.12.0. Can you confirm?
http://distro.ibiblio.org/fatdog/packag ... 6_64-1.txz

Greetings!

[stemsee]

There is a script which is backing up the /etc/resolv.conf file using pid number in the name, see pic .... resulting in a build up of resolv.conf-'pid number' files in /etc see pic

I haven't found the culprit yet!

[step]

build up of resolv.conf-'pid number' files in /etc see pic

The picture shows very close PIDs, so you're looking for a rapid fire sequence of the same script or command. In an unmodified Fatdog64-720 system - first-time booted with an empty savefile/savefolder - /etc/resolv.conf is updated by dhcpcd, which is (re)started each time a connection is brought down and up again. You can watch this happening by running htop and selecting "restart connection" from the wpa_gui tray icon right-click menu. You could also try to set a watch on resolv.conf with command inotifywait -m /etc/resolv.conf or make use of lsof to see which process is changing your file.

[kirk]

How small can a Fatdog64 installation partition be and still function? Either with save folder there or elsewhere?

We don't support "traditional" full installs. A Fatdog64 install is just copying the initrd and vmlinuz to the partition, so if you already have a boot loader installed it's just the size of those two files. What I usually do is make a one or two gigabyte ext3 partition for the intrd and kernel. Then put the savefolder on an ext4 partition. Some bootloaders (grub legacy and gub4dos) are slow loading from ext4.

[rufwoof]

What I usually do is make a one or two gigabyte ext3 partition for the intrd and kernel. Then put the savefolder on an ext4 partition. Some bootloaders (grub legacy and gub4dos) are slow loading from ext4.

I like the flexibility of ext3. Can be mounted as though a ext2 from the likes of BSD, or can be mounted as though a ext4 by Linux. That outweighs the benefit of running native ext4 in my book, so I format (Linux partitions) to ext3 exclusively.

[dr. Dan]

@kirk:
Thanks!

@rufwoof:
Also thanks for the knowledge.

Dan

[snayak]

I got setuptools-38.4.0.zip from https://pypi.python.org/pypi/setuptools. Installed it manually using command

Code: Select all

python setup.py install --record files.txt

Then I tried to install pip using easy_install.
It says

Code: Select all

ImportError: No module named extern

Tried different versions of setuptools, but failed. Tried installing extern 0.2.1 from https://pypi.python.org/pypi/extern/ manually and got the same error. Tried installing importlib 1.0.4 from https://pypi.python.org/pypi/importlib manually, same error.

Am I missing something?
Had not faced any such issue on Fatdog 611.


Sincerely,
Srinivas Nayak

[chiefengineer]

Has W10 on hard disk. Just want to leave it there right now. Bios allowed me to boot legacy USB. 710 works great. 720 lights the USB, screen blips a little, then all dark...the beta version got to the splashscreen then went dark forever. Tried in all USB slots with three different USB keys. Turned off TPM in bios no help.

Am I stuck on 710? Used both 710 and 720 to install to key through ISO file. The 720 keys work on other machines fine. What's different here? Have never entered the UEFI world (I just shut it off)...has it become necessary?