mklive.sh: running on foreign distros is broken (with static XBPS)

[JamiKettunen]

Using static XBPS from a foreign distro such as Ubuntu or Arch causes the process to get stuck (no new output without finishing script after 20 minutes) towards the end, specifically at cp -a "$ROOTFS"/* "$BUILDDIR"/tmp-rootfs/ of generate_squashfs(); here's the full set -x script log output: https://devspace.voidlinux.org/abby/paste/mklive-foreign-log.txt

The worst part is it runs an rm -rf command if you ^C to cancel the process (which was stuck) which also starts removing all stuff possible from under the host bind-mounted /sys, /proc and /dev, after which the system practically has to be rebooted due to missing /dev/null and others, not sure if this could even wipe some efivars possibly rendering host machines unbootable... Here's a short example:

+ error_out 130
+ umount_pseudofs
+ umount -R -f .../void-mklive/tmp.peM67GOz38/image/rootfs/sys
+ umount -R -f .../void-mklive/tmp.peM67GOz38/image/rootfs/dev
+ umount -R -f .../void-mklive/tmp.peM67GOz38/image/rootfs/proc
+ [ -d .../void-mklive/tmp.peM67GOz38 -a -z  ]
+ rm -rf .../void-mklive/tmp.peM67GOz38
rm: cannot remove '.../void-mklive/tmp.peM67GOz38/image/rootfs/sys/kernel/notes': Operation not permitted
...
rm: cannot remove '.../void-mklive/tmp.peM67GOz38/image/rootfs/dev/pts/ptmx': Operation not permitted
rm: cannot remove '.../void-mklive/tmp.peM67GOz38/tmp-rootfs': Device or resource busy
+ exit 130

TODO: Check how the output differs on a Void system...

[JamiKettunen]

A bit of progress, adding --lazy (-l for short) to the umount commands in umount_pseudofs helps, but now there's a few new problems:

1. "$VOIDHOSTDIR"/usr/bin/mksquashfs isn't found https://github.com/void-linux/void-mklive/blame/c9dbeed/mklive.sh.in#L271-L272, installing squashfstools on the foreign host system to provide mksquashfs doesn't help

[9/9] Generating squashfs image (xz) from rootfs...
./mklive.sh: 272: .../void-mklive/tmp.pgtMJbawV6/void-host/usr/bin/mksquashfs: not found
ERROR: Failed to generate squashfs image

2. It appears pseudofs dirs are umount'ed on the host after the (failed) execution of mklive.sh! here's a diff of mount command output before vs. after execution:

--- mounts.a
+++ mounts.b
@@ -1,23 +1,9 @@
 sysfs on /sys type sysfs (rw,nosuid,nodev,noexec,relatime)
 proc on /proc type proc (rw,nosuid,nodev,noexec,relatime)
 udev on /dev type devtmpfs (rw,nosuid,relatime,size=1939704k,nr_inodes=484926,mode=755,inode64)
-devpts on /dev/pts type devpts (rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000)
 tmpfs on /run type tmpfs (rw,nosuid,nodev,noexec,relatime,size=399708k,mode=755,inode64)
 /dev/mapper/data-root on / type ext4 (rw,noatime,errors=remount-ro)
-securityfs on /sys/kernel/security type securityfs (rw,nosuid,nodev,noexec,relatime)
-tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev,inode64)
 tmpfs on /run/lock type tmpfs (rw,nosuid,nodev,noexec,relatime,size=5120k,inode64)
-cgroup2 on /sys/fs/cgroup type cgroup2 (rw,nosuid,nodev,noexec,relatime,nsdelegate,memory_recursiveprot)
-pstore on /sys/fs/pstore type pstore (rw,nosuid,nodev,noexec,relatime)
-efivarfs on /sys/firmware/efi/efivars type efivarfs (rw,nosuid,nodev,noexec,relatime)
-bpf on /sys/fs/bpf type bpf (rw,nosuid,nodev,noexec,relatime,mode=700)
-systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=29,pgrp=1,timeout=0,minproto=5,maxproto=5,direct,pipe_ino=18870)
-mqueue on /dev/mqueue type mqueue (rw,nosuid,nodev,noexec,relatime)
-hugetlbfs on /dev/hugepages type hugetlbfs (rw,relatime,pagesize=2M)
-debugfs on /sys/kernel/debug type debugfs (rw,nosuid,nodev,noexec,relatime)
-tracefs on /sys/kernel/tracing type tracefs (rw,nosuid,nodev,noexec,relatime)
-fusectl on /sys/fs/fuse/connections type fusectl (rw,nosuid,nodev,noexec,relatime)
-configfs on /sys/kernel/config type configfs (rw,nosuid,nodev,noexec,relatime)
 none on /run/credentials/systemd-sysusers.service type ramfs (ro,nosuid,nodev,noexec,relatime,mode=700)
 /dev/vda1 on /boot/efi type vfat (rw,relatime,fmask=0077,dmask=0077,codepage=437,iocharset=iso8859-1,shortname=mixed,errors=remount-ro)
 tmpfs on /run/user/1000 type tmpfs (rw,nosuid,nodev,relatime,size=399704k,nr_inodes=99926,mode=700,uid=1000,gid=1000,inode64)

[Eloitor]

I have the same problem in bedrocklinux.

[classabbyamp]

is this still occurring since the possible fix of #365?

[bladedvox]

seems it's not completely resolved; umount_pseudofs is failing at the start of generate_squashfs (tested on an Ubuntu host):

umount /home/nilix/src/void-mklive/mklive-build.uCvf1/image/rootfs/sys/fs/cgroup: target is busy

[bladedvox]

using the solution mentioned in #414 does work, though!