With less than 130kB installed size it's a no-brainer and uuidgen is something we want to have on each Grml system available. Thanks: Raoul <firstname.lastname@example.org> for the feature request
We don't have log() available at that stage yet.
We no longer depend on the file /etc/grml/grml-live.conf as we keep system defaults in the program nowadays, so do not fail if file can not be read. Thanks: Evgeni Golov for reporting
As of commit ff80c12 ['Install "GRML" files into GRML/$GRML_NAME/'] the squashfs file is missing in the md5sum file, even though the squashfs file is the most important one to check for.
Nowadays even file-rc depends on insserv, so let's see how well this works out and be prepared...
It's a hard to debug problem as it doesn't seem to be reproducible in every OpenVZ environment and on the other hand makes the system unusable as soon as the bug is triggered. :( On the command line/in the logs you'll notice something like that before the system is becoming unuseable: | Setting up grub-common (1.99-22.1) ... | Setting up grub2-common (1.99-22.1) ... | Setting up grub-pc-bin (1.99-22.1) ... | Setting up grub-pc (1.99-22.1) ... | | Creating config file /etc/default/grub with new version | grub-probe: error: cannot find a device for / (is /dev mounted?). | grub-probe: error: cannot find a device for /boot (is /dev mounted?). | grub-probe: error: cannot find a device for /boot/grub (is /dev mounted?). To avoid grub-probe being executed during software installation let's dpkg-divert the binary during the FAI dirinstall process through the existing hook script instsoft.GRMLBASE. Then revert the dpkg-divert via scripts/GRMLBASE/98-clean-chroot before actually building the ISO again. While at it also divert update-grub in the same code path as well.
Useful for testing if multicast is working between two nodes, see e.g. http://pve.proxmox.com/wiki/Multicast_notes#test_if_multicast_is_working_between_two_nodes
When executing something like: | grml-live [...] -V B -q this doesn't skip the mksquashfs stage (and neither the chroot build stage of course) because the "B" can't be processed for obvious reasons. The user mistyped the "-B" here, but then we shouldn't even try to execute the command line, so just exit in such a case.
This is more user friendly than just displaying an error message
…n, see #677101) See http://bugs.debian.org/677101
This reverts commit 0533a72. "E: Method gave invalid 200 URI Start message" seems to be related to http.debian.net
Acked-by: Alexander Wirt <email@example.com>
The package is gone from Debian (RoQA; unmaintained, RC-buggy), see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=656590
We've reports that with TMPDIR being set the FAI scripts fail inside the chroot, so let's avoid that.
…y command dpkg v1.16.4 has a regression, failing the dpkg-query command to execute correctly. Because of that our daily builds fail, no need to force usage of the field width option, so let's just drop it. See Debian's BTS #676796