Dev1 Galaxy Forum / refractasnapshot-base build segfaults in chimaera

Re: refractasnapshot-base build segfaults in chimaera

dummy@example.com (nixer) — Fri, 27 Aug 2021 13:18:59 +0000

Well, I hope it is not something particular to my personal build. I doubt it. I wanted to let you know in case there was a problem somewhere with the function of refracta-tools and chimaera.

Refractasnapshot does work on chimaera if used on a nox system. I tested on a chimaera netinstall build using either the recent chimaera test installs, or a beowulf dist-upgrade of a stable netinstall. I don't think colord and sane is installed on that.

Thank you for your help.

Re: refractasnapshot-base build segfaults in chimaera

dummy@example.com (fsmithred) — Fri, 27 Aug 2021 12:57:17 +0000

Maintainer says he has no idea but will give it some thought. Also "colord-insane is a wrath upon us since the beginning."

I can't figure out what colord does. I have libcolord2 on the two systems I checked, but not colord. One of those systems is chimaera with runit, and I have made bootable isos from that system using refractasnapshot. Maybe removing colord will fix it.

Re: refractasnapshot-base build segfaults in chimaera

dummy@example.com (nixer) — Thu, 26 Aug 2021 19:36:14 +0000

Thanks for those commands, it worked fine. I emailed you a link to the text file of the error noted in log files. Lightdm does not start and the only specific problem that I see noted is the segfault with colord or sane file/package.

Also, to clarify this:

The chimaera snapshot iso boots in vbox in beowulf but not in chimaera?

No, the chimaera snapshot will not fully boot. Lightdm does not start. It does not matter if the physical host is booted into beowulf, chimaera, or daedalus, or ceres. It is the snapshot of the chimaera and daedalus install(s) that has the same segfault problem. A snapshot of the ceres system will boot fine. Sorry for any confusion.

Re: refractasnapshot-base build segfaults in chimaera

dummy@example.com (fsmithred) — Thu, 26 Aug 2021 18:35:33 +0000

if ssh server is installed in the snapshot, add the following to the qemu line:

-device e1000,netdev=user.0 -netdev user,id=user.0,hostfwd=tcp::5555-:22

And then you can ssh in with

ssh -p5555 user@localhost

where user is the actual username.

Note: in the qemu options above, 'user' does not get changed to the real username.

Re: refractasnapshot-base build segfaults in chimaera

dummy@example.com (nixer) — Thu, 26 Aug 2021 16:02:36 +0000

I installed qemu and booted the same iso file. It showed the same segfault message and continued booting for just a couple seconds past this point and it just stops. The last 3 lines of the screen text mentions successfully starting three programs: vboxdrv, VBoxNetFit, and VBoxNetAdp. It just stops at this point. I don't know the ip address to try to ssh into the VM to see any error log entry.

Is the first time I have used qemu. Thanks.

Will be glad to try something else. Will try the 'set-x' thing soon, when I make another snapshot.

I did try the 'set -x' code to the file. It post a little more code on the boot screen, but I could not find it in any other error log, like boot.log. I was not able to ssh into the VM but I did take a screenshot of the last bit of text, in case you want to see it.

Re: refractasnapshot-base build segfaults in chimaera

dummy@example.com (fsmithred) — Thu, 26 Aug 2021 15:59:58 +0000

Wow. I'm stumped. The chimaera snapshot iso boots in vbox in beowulf but not in chimaera?
Things to try:

Boot the iso in chimaera using qemu. You might need more that 1024mb of RAM.
qemu-system-x86_64 -m 1024 -cdrom snapshot.iso

Add 'set -x' to the top of /etc/init.d/saned to see if you get more information from the script.

There is no init script for colord, but there is a systemd service file. I don't know if the lack of an init script is the problem. Was there ever one for colord?

Not sure what to do with this. I'll ping jaromil. He's the maintainer of colord.

apt-file find colord-sane
colord: /usr/lib/colord/colord-sane       
colord: /usr/libexec/colord-sane

refractasnapshot-base build segfaults in chimaera

dummy@example.com (nixer) — Thu, 26 Aug 2021 13:51:55 +0000

I made a snapshot of my chimaera (dist-upgrade from beowulf) install and upon booting the iso in virtualbox, the boot segfaults with this error:

Aug 26 04:59:46 localhost kernel: [ 11.206762] colord-sane[2241]: segfault at 0 ip 00007f4734d14ad1 sp 00007ffcfecb03a8 error 4 in libc-2.31.so[7f4734bda000+14b000]
Aug 26 04:59:46 localhost kernel: [ 11.206785] Code: 84 00 00 00 00 00 0f 1f 00 31 c0 c5 f8 77 c3 66 2e 0f 1f 84 00 00 00 00 00 89 f9 48 89 fa c5 f9 ef c0 83 e1 3f 83 f9 20 77 1f fd 74 0f c5 fd d7 c1 85 c0 0f 85 df 00 00 00 48 83 c7 20 83 e1

I duplicated that this is repeatable, and I upgraded the same system (in a different partition) to daedalus and the problem exists there as well. This is a desktop system, with lightdm and xfce, and a nvidia graphics card - in case it is relevant.

The error message mentions both colord and sane, so I will include the installed package information for each. These are my installed packages in the chimaera install:

root@localhost:/home/user# apt policy *sane*
sane-utils:
Installed: 1.0.31-4.1
Candidate: 1.0.31-4.1
Version table:
*** 1.0.31-4.1 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
100 /var/lib/dpkg/status
xsane-common:
Installed: 0.999-10
Candidate: 0.999-10
Version table:
*** 0.999-10 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
500 http://deb.devuan.org/merged chimaera/main i386 Packages
100 /var/lib/dpkg/status
libsane-hpaio:
Installed: 3.21.2+dfsg1-2
Candidate: 3.21.2+dfsg1-2
Version table:
*** 3.21.2+dfsg1-2 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
100 /var/lib/dpkg/status
libsane-common:
Installed: 1.0.31-4.1
Candidate: 1.0.31-4.1
Version table:
*** 1.0.31-4.1 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
500 http://deb.devuan.org/merged chimaera/main i386 Packages
100 /var/lib/dpkg/status
xsane:
Installed: 0.999-10
Candidate: 0.999-10
Version table:
*** 0.999-10 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
100 /var/lib/dpkg/status
libsane1:
Installed: 1.0.31-4.1
Candidate: 1.0.31-4.1
Version table:
*** 1.0.31-4.1 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
100 /var/lib/dpkg/status

My installed colord packages:

root@localhost:/home/user# apt policy *colord*
libcolord2:
Installed: 1.4.5-3+devuan1
Candidate: 1.4.5-3+devuan1
Version table:
*** 1.4.5-3+devuan1 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
100 /var/lib/dpkg/status

colord:
Installed: 1.4.5-3+devuan1
Candidate: 1.4.5-3+devuan1
Version table:
*** 1.4.5-3+devuan1 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
100 /var/lib/dpkg/status

colord-data:
Installed: 1.4.5-3+devuan1
Candidate: 1.4.5-3+devuan1
Version table:
*** 1.4.5-3+devuan1 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
500 http://deb.devuan.org/merged chimaera/main i386 Packages
100 /var/lib/dpkg/status

root@localhost:/home/user# apt policy refractasnapshot-base*
refractasnapshot-base:
Installed: 10.2.12
Candidate: 10.2.12
Version table:
*** 10.2.12 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
500 http://deb.devuan.org/merged chimaera/main i386 Packages
100 /var/lib/dpkg/status

root@localhost:/home/user# apt policy refractainstaller-base*
refractainstaller-base:
Installed: 9.6.0
Candidate: 9.6.0
Version table:
*** 9.6.0 500
500 http://deb.devuan.org/merged chimaera/main amd64 Packages
500 http://deb.devuan.org/merged chimaera/main i386 Packages
100 /var/lib/dpkg/status

The system is up to date

root@localhost:/home/user# apt dist-upgrade
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Calculating upgrade... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

I don't know where to look on this, but the system(s) both boot fine without issue. It is just the snapshot build will not boot.
The refractasnspshot build boots fine in beowulf.

Edit: I will add this, in case it is helpful. The snapshot boots fine in ceres.

The *sane* packages in ceres are slightly newer, and different. But the libsane-hpaio package is the same.

root@localhost:/home/user# apt policy *sane*
sane-utils:
Installed: 1.0.32-4
Candidate: 1.0.32-4
Version table:
*** 1.0.32-4 100
100 /var/lib/dpkg/status
xsane-common:
Installed: 0.999-12
Candidate: 0.999-12
Version table:
*** 0.999-12 100
100 /var/lib/dpkg/status
libsane-hpaio:
Installed: 3.21.2+dfsg1-2
Candidate: 3.21.2+dfsg1-2
Version table:
*** 3.21.2+dfsg1-2 100
100 /var/lib/dpkg/status
libsane-common:
Installed: 1.0.32-4
Candidate: 1.0.32-4
Version table:
*** 1.0.32-4 100
100 /var/lib/dpkg/status
xsane:
Installed: 0.999-12
Candidate: 0.999-12
Version table:
*** 0.999-12 100
100 /var/lib/dpkg/status

libsane1:
Installed: 1.0.32-4
Candidate: 1.0.32-4
Version table:
*** 1.0.32-4 100
100 /var/lib/dpkg/status

The refracta packages are identical. and I only have the package colord installed in ceres. /end Edit