antiX live usb issue

[rs55]

@rs55, please run LUM again with fuse enabled. When you boot, please add these two options:

Code: Select all

checkmd5  bp=e

The first will check the integrity of the linuxfs file before we try to mount it. The 2nd will get you to a shell when the error occurs. In the shell run the command:

Code: Select all

dmesg | tail -n20

This should tell us why the mount failed. You could possible take a picture of the output if that is easy for you and it seems warranted.

If we are still getting fragmentation problems and we can't fix them then we may have to change our strategy and only use fuse when normal mounting fails. If I do give up then that will mean there will be some (rare) systems that LUM will not work on which is not a desirable outcome.

I want to first find out exactly what the problem is and then deal with it.

The error message from dmesg was SQUASHFS error: .... lz4 compression not supported.

[rs55]

also said:
"random: 7 urandom warning(s) missed due to ratelimiting" , after that lz4 compression error

[BitJam]

The error message from dmesg was SQUASHFS error: .... lz4 compression not supported.

Excellent! Thank you!

This is what I suspected. It is totally UNRELATED to fragmentation. The problem is the kernel does not support lz4 compression. I don't see how --force=nofuse would fix this problem if you start with the same iso file.

I've already made changes to live-remaster and live-kernel-updater to prevent this. We also need to make similar changes to the snapshot programs.

If you downloaded this iso file from us then please tell us so we can take it down at once.

I am very sorry that you keep bumping into problems. All I can say is that we are on the case and we are trying to fix them as best and as fast as we can.

[rs55]

The error message from dmesg was SQUASHFS error: .... lz4 compression not supported.

Excellent! Thank you!

This is what I suspected. It is totally UNRELATED to fragmentation. The problem is the kernel does not support lz4 compression. I don't see how --force=nofuse would fix this problem if you start with the same iso file.

I've already made changes to live-remaster and live-kernel-updater to prevent this. We also need to make similar changes to the snapshot programs.

If you downloaded this iso file from us then please tell us so we can take it down at once.

I am very sorry that you keep bumping into problems. All I can say is that we are on the case and we are trying to fix them as best and as fast as we can.

No problem at all. Thanks for working away at this. Important to make LUM perfect!!
By the way, I am now running kernel 4.9.0-8 and I created the iso and ran LUM under this kernel. No this was not a downloaded iso - this was an iso I created off the system.
This may not be relevant - but about a month or so ago - when I first started using MX, I had problems with lz4, which went away when I used lzo . But then more recently lz4 has been working fine. And it does work with nofuse.

[rs55]

Is the new LUM using a different version of squashfs ?

[BitJam]

Is the new LUM using a different version of squashfs ?

LUM does not use mksquashfs.

AFAIK, it is only used in snapshot, live-remaster, and our build-iso system. I've already added safe-guards to live-remaster and live-kernel-updater (but have not tested them yet due to a spate of emergencies) and I just sent an email to other devs suggesting the same change/fixes be made to the snapshot programs.

My fix is to check which forms of squashfs compression the current kernel supports and only offer the forms of compression that are supported.

[anticapitalista]

It does seem it is kernel related.

[rs55]

Is the new LUM using a different version of squashfs ?

LUM does not use mksquashfs.

AFAIK, it is only used in snapshot, live-remaster, and our build-iso system. I've already added safe-guards to live-remaster and live-kernel-updater (but have not tested them yet due to a spate of emergencies) and I just sent an email to other devs suggesting the same change/fixes be made to the snapshot programs.

My fix is to check which forms of squashfs compression the current kernel supports and only offer the forms of compression that are supported.

I did another experiment. I created a new iso file using lzo compression. Then created a live usb with the new lum. That failed as well .

[fehlix]

I did another experiment. I created a new iso file using lzo compression. Then created a live usb with the new lum. That failed as well .

@rs55, please do report exactly what you did which program you used, and under what condition it failed.
what kernel. You already reported today and I asked which ISO, you used, and you didn't answer.
It doesn't help much if just report in an emergency style something failed with a certain option used.
If you don't provide exact information, you just creating confusion, as partly seen within this thread, already.
Thanks
fehlix

[rs55]

And - just now, I created a usb from the command line with force=nofuse. Same iso, same everything. It worked like a charm.
Sooo - this fuse is the culprit.
I am willing to bet I am not alone in this - because I am not using anything weird, or home brewed. Iam using a plain old standard thinkpad from when thinkpads were great.
You have perhaps not seen many complaints yet - because with the old LUM ( in the repo until this morning) , fuseiso was never actually getting invoked, because the standard installation installs fuseiso9660.
But - with todays' update - I bet you see a lot of folks who have problems with the lum.