Files copied from generating system
We can't really predict the behavior of or support initrd's that end up with whatever customized config files are copied from the system, but options for dealing with this problem are not clear. Any ideas?
modules.d/95nfs/install:
Quoted 2 lines
/etc/nsswitch.conf/etc/idmapd.conf
These could possibly be edited and cause NFS to behave unexpectedly in the initrd.
(Harald and I agree we should write the needed user to /etc/passwd during dracut generation instead of copying the /etc/passwd. I'm working on this.)
modules.d/90kernel-modules/install:
Quoted 2 lines
[ -f
/etc/modprobe.conf] &&dracut_install/etc/modprobe.confdracut_install$(find/etc/modprobe.d/-type f -name '*.conf')
Some of these are shipped by the distribution and important, but others can be edited and cause the initrd to behave differently.
modules.d/90mdraid/install:
Quoted 4 lines
#!/bin/bash
dracut_installmdadm
inst/etc/passwd
inst/etc/group
With agreement from Harald we're writing these during dracut generation instead of copying.
Quoted 8 lines
instmods =drivers/md
inst_rules"$moddir/61-mdadm.rules"
[ -f/etc/mdadm/mdadm.conf] && inst/etc/mdadm/mdadm.conf/etc/mdadm.conf[ -f/etc/mdadm.conf] && inst/etc/mdadm.conf
if [ -x/sbin/mdmon] ; then
dracut_installmdmon
inst_hookpre-pivot 30 "$moddir/mdmon-pre-pivot.sh"
fi
mdadm.conf is another file that can influence the behavior of a system.
Shouldn't this be --hostonly?
Is dracut mdraid module usable at all without mdadm.conf? I've used many RAID systems without mdadm.conf, although I don't understand why some systems seemingly need it while others don't.
modules.d/95udev-rules/install:
Quoted 11 lines
#!/bin/bash
# FIXME: would be nice if we didn't have to know which rules to grab.... # ultimately,/lib/initramfs/rules.dor somesuch which includes links/copies# of the rules we want so that we just copy those in would be best dracut_install udevd udevadm /lib*/udev/*_id /etc/udev/udev.conf /etc/group dracut_install basename inst_rules 50-udev-default.rules 60-persistent-storage.rules \ 61-persistent-storage-edd.rules 80-drivers.rules 95-udev-late.rules #Some debian udev rules are named differently inst_rules 50-udev.rules 95-late.rules
Quoted 5 lines
if ldd $(find_binary udevd) |grep -q
/lib64/libc; then
dracut_install/lib64/libnss_files*
else
dracut_install/lib/libnss_files*
fi
udev.conf could be another source of unexpected behavior?
Warren Togami
[email protected]
--
To unsubscribe from this list: send the line "unsubscribe initramfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 10 lines · Warren Togami
We can't really predict the behavior of or support initrd's that end up with whatever customized config files are copied from the system, but options for dealing with this problem are not clear. Any ideas?
modules.d/95nfs/install:
Quoted 2 lines
/etc/nsswitch.conf/etc/idmapd.confThese could possibly be edited and cause NFS to behave unexpectedly in the initrd.
rpc.idmapd and /etc/idmapd.conf can probably go away, since it was decreed that we only have to mount the root filesystem and the distro has to worry about getting ID mapping working before using a non-root user.
nsswitch.conf is more difficult, but if you are relying on DNS, NIS, or LDAP to find your network root, this is a concern you should probably be aware of.
--
To unsubscribe from this list: send the line "unsubscribe initramfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html