| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
| |
These decisions will now be made in anaconda proper. There's no need to
involve three separate processes in the discussion.
|
| |
|
|
|
|
|
|
| |
Previously killall5 omitted the init's session (which includes the anaconda
process group) and anaconda and everything spawned from there (including
the vnc server) survived both the killall5 calls.
Related: rhbz#679397
|
| |
|
|
|
| |
Without this the termination procedure just prints errors about how it can
not find the binaries.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
We switched to doing umounts beforehand and also to using lazy
umount to prevent hangs due to having killed NetworkManager prior
to unmounting NFS media. (commit a1c759a9524e003)
With this we perform terminations first, but we use killall5's
ability to omit certain pids from the list of processes to kill.
We kill everything except mdmon (this bug), NetworkManager, and
dhclient (for the NFS unmount thing, bz#463959). We don't need
to do lazy unmounting anymore, either.
Resolves: rhbz#604614
|
| |
|
|
|
|
|
|
|
| |
When deciding if we need to mv /lib64_old to /lib64 or /lib_old to /lib
after unmounting /mnt/runtime, lstat must be used as /lib64 if present
is a link to a no longer existing dir, so using regular stat to test for
/lib64 will fail as regular stat follows links.
Related: rhbz#604614
|
| |
|
|
|
|
|
|
|
| |
Ignore errors upon restoring /lib and /usr after unmounting filesystems,
otherwise these errors always get reported when rebooting before we've
reached stage2. Instead add proper error checking to the execution of
mdadm.
Related: rhbz#598222
|
| |
|
|
|
|
|
|
|
|
|
| |
When moving to stage2 we rename the initrd /lib to /lib_old and add
a symlink from /lib to /mnt/runtime/lib. When shutting down we unmount
filesystems and then after unmount try to run mdadm to wait for
any Intel BIOS RAID arrays to have their metadata updated that they
were cleanly shutdown to avoid a resync of the entire array on the next boot.
But at this point /lib is a dangling symlink, so no elf interpreter gets
found (nor libc) and the mdadm call we do fails, leaving the array unclean.
|
| |
|
|
|
|
|
|
| |
Stop the kernel correctly on HALT (so that we dont see 'kernel panic, they
killed init' on i386 and on s390 one gets easy access to manual IPL).
Introduce a new reboot method that does allow us to see the backtrace and
doesn't scroll the screen up with useless unmount info.
|
| | |
|
| |
|
|
|
|
|
|
| |
Do not install a handler for SIGINT on s390 since you can't really press
Ctrl+Alt+Del there. Only handle SIGUSR1 and SIGUSR2 as handled in
linuxrc.s390 (which comes from loader which comes from iutil.py).
Always make sure we unmount filesystems and kill off processes, then
reboot if that's requested or exit and display the halt message.
|
| |
|
|
|
| |
All previous comments about test mode being untested apply to loader, but
doubly so.
|
| |
|
|
|
|
|
|
|
| |
The support for boot option "nokill" in shutdown by
ade8105a8515afee24220fd0cc1199e30c1c3142 did not adapt the
standalone version of shutdown (#ifdef AS_SHUTDOWN) so doKill
was always 0. Introduce a new command line option "--nokill" and
otherwise default to doKill=1. While we're at it also provide
a command line option "-P" for poweroff.
|
| |
|
|
|
|
|
| |
With the new external metadata mdraid support we need to wait for raid
sets to become clean (iow for mdmon to have finished updating the metadata)
before rebooting. Otherwise we get a full array sync on the first boot after
installation.
|
| |
|
|
|
|
| |
This test causes it not to reboot if you don't specify anything about
rebooting in kickstart or on the command line. It also treats /all/
installs with "nokill" specified as "halt" installs.
|
| |
|
|
|
| |
Syntax of shutDown() function changed recently, so make sure we call it
correctly.
|
| |
|
|
|
|
| |
The new reboot code didn't take into account the fact that there's a
different sigint handler while anaconda is running vs when we're done
installing. This patch fixes that bad assumption.
|
| |
|
|
|
|
| |
Rework shutDown() so that we don't do unmounts until we're going to
reboot or power off the machine. Also some minor cleanups like taking
out all the negative-phrasing of "noKill" in the code itself.
|
| |
|
|
|
|
|
| |
This fixes a problem where we'd kill NM, then try to unmount our NFS
installation source. The result was a hang on the unmount and anaconda
never rebooting. Instead, unmount before killing processes and do the
unmounts lazily since /mnt/stage2 is likely to still be busy.
|
| |
|
