Stuff that has not yet gone into the official build.
Post a reply

Re: Alternative usb installation method, part 2

Mon Apr 07, 2014 8:53 pm

I tried it again after adding '--directory' and it worked correctly. The directions are correct - select (highlight) either /live in the root of the device or <dirname>/live, where dirname is the name of the directory that contains the image you want to replace on a multiboot usb. I successfully replaced the first image (in /live) and the additional image (in /sid/live).

Re: Alternative usb installation method, part 2

Sat Apr 12, 2014 10:32 pm

Uploaded a fixed version of refracta2usb at sourceforge. In addition to the directory-selection bug, I added pv to the deps and use it with the dd command in mkloopback. Found myself staring at a silent terminal for a long time when I made a large loopback file.

Re: Alternative usb installation method, part 2

Sun Apr 13, 2014 4:46 am

I'm happy that you've added pv.
Is this the same "pv" as ---} http://www.ivarch.com/programs/pv.shtml
If it is, are we missing any remarkable bugfixes in the current debian package?
Looks like the authors have patched several debian-reported bugs since v12.0.1
(which is the version, from 2010, contained in the debian repos)

What led you to choose pv instead of "pipemeter" or "bar"? Is pv easier to use, is smaller or ???

(I'm not a fan of the "gui" versions of the snapshot and 2usb scripts but)
in both the cli and gui versions, there are a few junctures where user is left wondering "did it hang? what is it doing?"
Yad progressbar is (still) missing for some of these, in the refracta as well as antix scripts...
...and pv (or bar or pipemeter) feedback of the progress would be a welcome addition to the cli versions.

Ah, I guess the above feedback is applicable to Makulu's (essentially unchanged) implementation of the scripts as well.

Re: Alternative usb installation method, part 2

Sun Apr 13, 2014 10:16 am

+1 for pv

Re: Alternative usb installation method, part 2

Sun Apr 13, 2014 11:33 am

Yeah, that's the one. I chose it on dzz's recommendation, and I didn't choose it over the others, because I didn't know they exist. I'll take a look.

BTW, the yad progress bar seems to be missing even in places where it's in the code. Can't tell you how many times I've changed the check_size function to get that progress bar to show, got it to work, and then when I run the program later, the progress bar doesn't show up. (It shows up for less than a tenth of a second at the end of the transfer.)

Huh. Maybe I need another cup of coffe. This is not working for me. Here's the command from the script:
Code:
dd if=/dev/zero | pv | of=${LOOP_FILENAME} bs=1M count=${SIZE}


Here's what I'm running to test it. The second line is the output. There is no output file. I also tried it with the dd options before the input file instead of after the output file. I'm sure I did this before adding pv to the script, and it worked.
Code:
dd if=/dev/zero | pv | of=testloop bs=1M count=1000
   0B 0:00:00 [   0B/s] [<=>


This does work for me (today, right now.)
Code:
dd bs=1M count=1000 if=/dev/zero | bar -of testloop


OK, the dd command with pv in the script does not work. I'm pulling down the deb file. Not sure what's happening after that.

Re: Alternative usb installation method, part 2

Sun Apr 13, 2014 1:35 pm

Code:
SIZE=50

# no progress bar without this if IO redirection to a log
# exec 2>&1

:~$ dd 'if=/dev/zero'  bs=1024 count=${SIZE}000 |pv -fprt -s ${SIZE}m |dd of=loopfile bs=1024 count=${SIZE}000

50000+0 records in
50000+0 records out
51200000 bytes (51 MB) copied, 0.634875 s, 80.6 MB/s
0:00:00 [76.7MB/s] [====================================================>  ] 97%
50000+0 records in
50000+0 records out
51200000 bytes (51 MB) copied, 0.636096 s, 80.5 MB/s


was just observed here. A 50MB file was written. Method and options from old notes.

Re: Alternative usb installation method, part 2

Sun Apr 13, 2014 6:42 pm

That works, thanks. I forgot about the redirect. I like the output better without the '-fprt' because it shows a count of total MB copied. Also added pv to mke2fs, It just counts the time, but at least something is happening.

Uploaded corrected refracta2usb-0.9.6.deb to sourceforge:
http://sourceforge.net/projects/refract ... b/download

Re: Alternative usb installation method, part 2

Mon Apr 14, 2014 1:43 am

+1 for pv means only, that's what I found and got to know (a little) that was seen to work..

Thanks for the updated deb!

BTW while on the case of alternatives, watch out for pmount.. it's unmaintained for 3 years and might disappear. For now, 0.9.99-alpha-1 in experimental (can handle loopfiles including luks) is good for wheezy and sid

Re: Alternative usb installation method, part 2

Mon Apr 14, 2014 7:58 am

dzz wrote:while on the case of alternatives, watch out for pmount.. it's unmaintained for 3 years and might disappear.
As a pmount alternative, check out the udevil package provided by the spacefm author
http://ignorantguru.github.io/udevil/

Re: Alternative usb installation method, part 2

Mon Apr 14, 2014 12:58 pm

just tried udevil.. it's similar to pmount. I will use it.

Has a plain text config file like pmount (editable to allow internals). It's in backports, jessie, sid but not wheezy main. 499 kB installed-size, no extra deps needed for my current Wheezy. Not DE-dependant (is cli, should be good for non-root scripts). Has a maintainer. No serious bug reports.

It does squashfs and iso but unfortunately not luks

Example usage (it makes the mountpoint if does not exist):
Code:
:~$ udevil mount /dev/sdc1 /media/zmount
Mounted /dev/sdc1 at /media/zmount

If no mountpoint specified it will be /media/disk_label
Post a reply