Win7x64 : Golden Image : partitions problems ... again

george1421

@Jonathan-Cool I would do a debug deploy. Then at the command prompt ensure that both fog.upload and funcs.sh have been properly patched.

Quazz

@Jonathan-Cool Ok, so the label should match now at the very least.

Did you recapture the image since implementing the latest sed?

If so, what do your d1 files look like for this image?

Jonathan Cool

Hi,
I’m back … i typed the same commands on client during a DL Debug Task … : same results : (with SSH)

[Fri Sep 28 root@fogclient ~]# fog
 * Running post init scripts.........................Done
 * Press [Enter] key to continue
^C
[Fri Sep 28 root@fogclient ~]# grep Vv /usr/share/fog/lib/funcs.sh 
            if [[ $label =~ [Rr][Ee][Cc][Oo][Vv][Ee][Rr][Yy] ]]; then
            if [[ $label =~ [Rr][Éé][Ss][Ee][Rr][Vv][Éé] || $label =~ [Rr][Éé][Ss][Éé][Rr][Vv][Éé] ]]; then
[Fri Sep 28 root@fogclient ~]# grep Vv /bin/fog.upload 
                            *[Rr][Ee][Cc][Oo][Vv][Ee][Rr][Yy]*|*[Rr][Éé][Ss][Ee][Rr][Vv][Éé]*)
[Fri Sep 28 root@fogclient ~]# blkid -po udev /dev/sda1 | grep LABEL
ID_FS_LABEL=Réservé_au_système
ID_FS_LABEL_ENC=Réservé\x20au\x20système
[Fri Sep 28 root@fogclient ~]#

The patch seem to be applied as you can see … " if [[ $label =~ [Rr][Éé][Ss][Ee][Rr][Vv][Éé]"

@Quazz Sorry, i’m not sure if i understood the question about this file … i take a screenshot of the folder of my Golden Img :
https://drive.google.com/file/d/1iN8YbEsgOsCSu622R9qG4JzRE9xjXo1B/view?usp=sharinga
But, yes, i recapture the image before lastest sed

Quazz

@Jonathan-Cool Can you share the contents of d1.fixed_size_partitions d1.partitions and d1.minimum.partitions

Jonathan Cool

Yes i can !

d1.fixed_size_partitions :

:2

d1.partitions :

label: dos
label-id: 0xc91a5cff
device: /dev/sda
unit: sectors

/dev/sda1 : start=        2048, size=      204800, type=7, bootable
/dev/sda2 : start=      206848, size=   335335424, type=7

d1.minimum.partitions :

label: dos
label-id: 0xc91a5cff
device: /dev/sda
unit: sectors

/dev/sda1 : start=        2048, size=       52316, type=7, bootable
/dev/sda2 : start=      206848, size=   335335424, type=7

What do you think about that ?

Quazz

@Jonathan-Cool It’s very strange. The regex should definitely match now, yet it marks the wrong partition as fixed_size for some reason.

I’ve looked over the code and can’t immediatly find anything desperately wrong or incorrect for the time being.

The strange thing, imo, is that it marks the second partition as fixed, whereas we’d expect it to mark none as fixed if it were to fail (unless I’m missing something in the code, but don’t think so)

Jonathan Cool

Maybe i choose wrong parameters when i created my VM with Virtual BOX the first time ? (details on the first topic) ?

Quazz

@Jonathan-Cool Did you switch to a fixed size as I suggested in the previous thread?

Though, I don’t see how that would cause this exact problem, to be honest.

It’s very weird that you get :2 in fixed_size_partitions as I get an empty file on my W7 French installs instead. Did you assign a label to the drive per chance?

It seems like our efforts haven’t quite made the difference I hoped yet. The regex matches in my bash tests, but doesn’t seem to match in FOS in a real world scenario. I’m rather baffled as to why that is, though.

Jonathan Cool

@Quazz said in Win7x64 : Golden Image : partitions problems ... again:
Did you assign a label to the drive per chance?

Yes, my drive is named “System” into my Golden Image … https://drive.google.com/file/d/0B48PuYLEHxQLUEU5clJOZC1Vb3c/view

Maybe this name is confusing FOG ?

Quazz

@Jonathan-Cool Ideally it shouldn’t cause any problems, but because the label system seems to be having some issues, it might be a good idea to remove it before capture at least.

That should leave you with an empty fixed_size_partitions file I think…

Still doesn’t answer why the system reserved isn’t matching, it doesn’t really make any sense to me off the top of my head.

Jonathan Cool

I will try to remove this label on the C drive and recapture the image … and try a new download on the optiplex 7050.

I will be back !

Quazz

@Jonathan-Cool My “fuck this shit” solution atm is the following if you’d like to try it

Note that this is experimental and can cause issues on systems with odd labels so use at own risk… but

sed -i -e "s#\\[Rr\\]\\[Ee\\]\\[Ss\\]\\[Ee\\]\\[Rr\\]\\[Vv\\]\\[Ee\\]\\[Dd\\]#[Rr]*[Ss][Ee][Rr][Vv]#gi" /bin/fog.upload
sed -i -e "s#\\[Rr\\]\\[Ee\\]\\[Ss\\]\\[Ee\\]\\[Rr\\]\\[Vv\\]\\[Ee\\]\\[Dd\\]#[Rr]*[Ss][Ee][Rr][Vv]*#gi" /usr/share/fog/lib/funcs.sh

This seems to allow the regex to match at least, of course it’s far less precise now. Still don’t know why it hates é so much, but that seems to be the issue unfortunately.

Should also work on English and Dutch systems I think.

Quazz

So after some testing, I think the main problem was the case pattern matching not being able to match for some reason in FOS. (works fine on my server, however)

I’ve updated my commit to change it to the if pattern matching instead (which I already tested actually works in FOS).

Don’t know when/if it gets merged, but that should resolve issues for French Windows installs at least.

Pull request here: https://github.com/FOGProject/fos/pull/13

Jonathan Cool

@Quazz
Hi.
I can try to remove my C label partition and see the result or it’s not necessary anymore ?
Same question for the last “sed” you proposed …

Maybe my best option is waiting the fix in a next release … ?

Quazz

@Jonathan-Cool I’d recommend trying both and seeing if it resolves your problem.

I don’t know when/if the PR will get accepted, could be a while.

updated the sed because the escapes got removed again.

Sebastian Roth

@Quazz Thanks heaps for testing and your pull request. Really weird that it was that hard to figure out for the French accents… Looks pretty good to me although I can’t test it.

@Jonathan-Cool I will see if I can manage to build and upload fresh init files in the next couple of days. This way you wouldn’t need to use the sed stuff.

Jonathan Cool

@Quazz said in Win7x64 : Golden Image : partitions problems ... again:

@Jonathan-Cool My “fuck this shit” solution atm is the following if you’d like to try it

Note that this is experimental and can cause issues on systems with odd labels so use at own risk… but
sed -i -e "s#\\[Rr\\]\\[Ee\\]\\[Ss\\]\\[Ee\\]\\[Rr\\]\\[Vv\\]\\[Ee\\]\\[Dd\\]#[Rr]*[Ss][Ee][Rr][Vv]#gi" /bin/fog.upload
sed -i -e "s#\\[Rr\\]\\[Ee\\]\\[Ss\\]\\[Ee\\]\\[Rr\\]\\[Vv\\]\\[Ee\\]\\[Dd\\]#[Rr]*[Ss][Ee][Rr][Vv]*#gi" /usr/share/fog/lib/funcs.sh

In the second line, we can see a “*”, is-it a mistake or not ? (line concerning funcs.sh)

After your answer, i will try :

to update my fog.postint
remove my C label

Quazz

@Jonathan-Cool It’s not a mistake, although you could write the line differently since that’s an if regex pattern match, which does work.

* means any character in pattern matching, which includes é of course

It worked for my images anyway, so hopefully you can reproduce those results.

Quazz

@Sebastian-Roth Thanks, I probably should have looked into it earlier, but don’t always have the time for such things.

What are your thoughts on https://forums.fogproject.org/topic/12428/win7x64-golden-image-partitions-problems-again/29 as an alternative system, by the way?

Potentially we could combine multiple systems to increase accuracy and reliability, too.

Jonathan Cool

Today, i’m the very happy french of the world … 'cause it’s WORKS !
I tried on o7010, o5050 and o7050 and it’s works !
Wahouuuuuuuuuuuuuuuu
I just want to say many thanks to you both and i think, it’s not enough … !

Long live to FOG Project

fog.postinit :

sed -i -e "s#\[Rr\]\[Ee\]\[Ss\]\[Ee\]\[Rr\]\[Vv\]\[Ee\]\[Dd\]#[Rr]*[Ss][Ee][Rr][Vv]#gi" /bin/fog.upload
sed -i -e "s#\[Rr\]\[Ee\]\[Ss\]\[Ee\]\[Rr\]\[Vv\]\[Ee\]\[Dd\]#[Rr]*[Ss][Ee][Rr][Vv]*#gi" /usr/share/fog/lib/funcs.sh

Label on C drive removed on the Golden image.

Happy Face

Solved !!

Win7x64 : Golden Image : partitions problems ... again

167

12.0k

17.3k

155.2k