openrc support thread

[Januszzz]

Edit by nixnut: split off from [NEWS] OpenRC and baselayout 2 will reach ~arch soon

openrc doesn't compile on uclibc experimental stage (stage3-x86-uclibc-2008.0_beta1.tar.bz2). How to reproduce:

Code: Select all

mkdir /custom && cd /custom
wget http://gentoo.po.opole.pl/releases/x86/2008.0_beta1/experimental/stage3-x86-uclibc-2008.0_beta1.tar.bz2
tar -xjvpf stage3-x86-uclibc-2008.0_beta1.tar.bz2
cd usr/
wget http://gentoo.po.opole.pl/snapshots/portage-latest.tar.bz2
tar -xjvpf portage-latest.tar.bz2
mount -t proc none /custom/proc/
mount -o bind /dev/ /custom/dev/
cp /etc/resolv.conf /custom/etc/

then prepare make.conf with this:

Code: Select all

 CFLAGS="-Os -mtune=i386 -pipe"
CXXFLAGS="-Os -mtune=i386 -pipe"
LDFLAGS="-Wl,-O1"
CHOST="i386-gentoo-linux-uclibc"
MAKEOPTS="-j12"
CLEAN_DELAY="0"
USE="ssl ipv6 unicode -cracklib"
FEATURES="parallel-fetch buildpkg"

then update the chroot, make new useflags:

Code: Select all

emerge -Du world && emerge -N world

the latter failed, so I did in meantime:

Code: Select all

FEATURES=-sandbox emerge coreutils

then unmask openrc:

Code: Select all

echo "sys-apps/baselayout" > /etc/portage/package.unmask
echo "sys-apps/baselayout" > /etc/portage/package.keywords
echo "sys-apps/openrc" >> /etc/portage/package.keywords
echo "sys-apps/openrc" >> /etc/portage/package.unmask

unmerge udev:

Code: Select all

emerge unmerge udev

Finally, emerge baselayout:

Code: Select all

emerge baselayout

and it dies with this: http://uni.opole.pl/janusz/log

Any hints?

[UberLord]

You should update to uclibc-0.9.29 or update your sys/queue.h header.

[Januszzz]

hmm, any chances newer uclibc is in Portage? if openrc hits ~arch uclibc should also.

[UberLord]

I would file a bug asking for that

[albright]

Yes, it is faster but I have a small problem.

shutdown -F does not seem to work (I have
fstab set to NOT check the drive periodically,
since, well, have you ever started your computer
to give a presentation and fsck starts ...)

Also, the silent splash screen mode doesn't show
the progress bar or give the F2 switch option (and
nothing happens when F2 is pressed)

Any advice appreciated ...

[UberLord]

The shutdown -f problem has already been reported and fixed in the git repo. Will be in openrc-0.2.2 or later.

[risa2000]

After reading the migration guide, I wonder what is the reason behind decision to encode names and versions into parameter names. E.g:

Code: Select all

# For 2.6.23-gentoo-r5, pass video_br=2 to cx88_dvb
module_cx88_dvb_args_2_6_23_gentoo_r5="video_br=2"
# For 2.6.x series kernels, always pass vendor and product
module_usbserial_args_2_6="vendor=0x1410 product=0x2110"
# Always pass debug to ieee1394
module_ieee1394_args="debug" 

I see module name mixed with uname (where some characters like dots and hyphen are converted to underscore).
Why not use some hierarchical structure (like INI file) instead?

Also I believe the module name and its (optional) parameters should not be divided into two lexically separated information. Either I give the module (and eventually its parameters) or I do not. Or is it foreseen that it will be possible to provide parameters for modules, without explicitly specifying the modules?

I would probably like more something like this:

Code: Select all

[*]
ieee1394 debug
[2.6]
usbserial vendor=0x1410 product=0x2110
[2.6.23-gentoo-r5]
cx88_dvb video_br=2

[virdiq]

A sample /etc/timezone doesn't seem to be included. Does anyone know what the format of this file should be (or whether I need another package which provides this file)? Thanks.

[Kollin]

A sample /etc/timezone doesn't seem to be included. Does anyone know what the format of this file should be (or whether I need another package which provides this file)? Thanks.

Yeah it`s missing

[s4e8]

It's not fun with the new goody. I'am happy I choose to upgrade the chrooted 32bit system first.

>>> Completed installing openrc-0.2.1-r2 into /tmp/portage/sys-apps/openrc-0.2.1-r2/image/

ecompressdir: bzip2 -9 /usr/share/man
strip: i686-pc-linux-gnu-strip --strip-unneeded -R .comment
sbin/rc
lib/librc.so.1
lib/libeinfo.so.1
usr/lib/librc.a
usr/lib/libeinfo.a
making executable: lib/libeinfo.so.1
making executable: lib/librc.so.1

* QA Notice: Missing gen_usr_ldscript for libeinfo.so

* QA Notice: Missing gen_usr_ldscript for librc.so
*
* ERROR: sys-apps/openrc-0.2.1-r2 failed.
* Call stack:
* misc-functions.sh, line 617: Called install_qa_check
* misc-functions.sh, line 259: Called die
* The specific snippet of code:
* [[ ${abort} == "yes" ]] && die "add those ldscripts"
* The die message:
* add those ldscripts

[virdiq]

A sample /etc/timezone doesn't seem to be included. Does anyone know what the format of this file should be (or whether I need another package which provides this file)? Thanks.

Yeah it`s missing

It seems that /etc/timezone will be created from /etc/conf.d/clock as follows:

Code: Select all

	# upgrade timezone file
	if [[ -e ${ROOT}/etc/conf.d/clock && ! -e ${ROOT}/etc/timezone ]] ; then
		(
		source "${ROOT}"/etc/conf.d/clock
		[[ -n ${TIMEZONE} ]] && echo "${TIMEZONE}" > "${ROOT}"/etc/timezone
		)
	fi

(from http://sources.gentoo.org/viewcvs.py/ge ... iew=markup)

But this upgrade didn't appear to work for me?

[Kabuto]

Eeep! I lost net.eth0. Am I missing something?

UPDATE: I just created a net.eth0 linked to /lib/rc/sh/net.sh like my net.lo was and that worked.

[ocin]

migrated and now my net.eth0 is gone and /etc/conf.d/net is empty. network does not work anymore. how to solve this (since /etc/runlevels/default/net.eh0 points to /etc/init.d/net.eth0 which is non-existant - how to get that file back)?

[lorenb]

Perhaps the OpenRC guide should mention that it will delete your /etc/init.d/net.eth0 (and similar files). The upgrade broke my networking that way. That was an unpleasent surprise.

I noticed a bunch of other services broke on this box too but perhaps it's related to the network issue above. Trying to debug now.

EDIT: Found it also deleted my /etc/conf.d/net file too. Once I created a new one and made a symlink for net.eth0 everything was fixed.

[virdiq]

Eeep! I lost net.eth0. Am I missing something?

UPDATE: I just created a net.eth0 linked to /lib/rc/sh/net.sh like my net.lo was and that worked.

I found the answer to why net.eth0/etc have disappeared (it'd be good to get networking changes added to the migration guide).

Also from the openrc ebuild, here is the upgrade procedure it attempts:

Code: Select all

	# force net init.d scripts into symlinks
	for f in $(ls "${ROOT}"/etc/init.d/net.*) ; do
		if [[ ! -L ${f} ]] ; then
			elog "Moved net service '${f##*/}' to '${f##*/}.openrc.bak' to force a symlink."
			elog "You should delete '${f##*/}.openrc.bak' if you don't need it."
			mv "${f}" "${f}.openrc.bak"
			ln -snf net.lo "${f}"
		fi
	done

All existing net.* links appear to be removed in the upgrade. If you had an actual file called 'net.eth0' it'll back it up and create a link by the same name to net.lo (a symlink to /lib/rc/sh/net.sh).

[ConiKost]

For me, the upgrade worked perfectly fine.

May I ask, for what is the ncurses and pam use flag good in openrc?

[zach9824]

Major issues with networking here as well.

I use a bridge setup for virtualbox. The dependencies are no longer working and it it will not create the br0 device on boot. Manually it works. Has the syntax of /etc/conf.d/net changed?

Code: Select all

##Bridge##
#the interface is really replaced by br0 for outgoing, so if you had dhcp here, you'll use dhcp in br0
config_eth0=( "null" )

#this specifies that interface vbox0 will be a tap interface and use the tunctl command to generate itself
tuntap_vbox0="tap"

#ALL interfaces part of a bridge should be null, otherwise oddities may occur.
config_vbox0=( "null")

# you can specify an owner of the interface if you want to run virtual box as a non root user
tunctl_vbox0="-u zach"

#If you'd prefer a dynamic ip address for the machine, use
config_br0=( "dhcp" )
#dhcp_br0="nodns nontp nonis"

#this specifies the bridging information
bridge_br0="eth0 vbox0"

depend_br0() {
rc_after_br0="net.eth0 net.vbox0"
}

brctl_br0=( "setfd 0" "stp on" )

[virdiq]

Code: Select all

depend_br0() {
rc_after_br0="net.eth0 net.vbox0"
}

net.eth0 and net.vbox0 won't exist after the upgrade (which is why those dependencies are failing). I imagine you'll need to create new symlinks named 'net.eth0' and 'net.vbox0', each with a target of 'net.lo'.

Check out the Gentoo handbook section on networking for more information on what you need to do.

Example:
cd /etc/init.d
ln -s net.lo net.eth0
rc-update add net.eth0 default

[IvanMajhen]

Works great for me. What a fast boot.

Splash doesn't work. Console decoration works, but there is no progress bar or icons when booting. Worked before upgrade.
Also, is it possible to automatically start dnsmasq after net.wlan0 or net.eth0 is connected?
For example, if net.wlan0 is stopped then dnsmasq also stops. But if i start net.wlan0 again, dnsmasq doesn't start.

[Kollin]

Code: Select all

depend_br0() {
rc_after_br0="net.eth0 net.vbox0"
}

net.eth0 and net.vbox0 won't exist after the upgrade (which is why those dependencies are failing). I imagine you'll need to create new symlinks named 'net.eth0' and 'net.vbox0', each with a target of 'net.lo'.

Check out the Gentoo handbook section on networking for more information on what you need to do.

Yeah just created the net.eth0 link again after reboot

[zach9824]

Still was not working until I changed the depend as follows:

Code: Select all

depend_br0() 
{
         need net.eth0
         need net.vbox0
}

Now Everything works great!

I also just noticed the irqbalance is no longer working, I will check if there is already a bug open:

Code: Select all

/etc/init.d/irqbalance start
 * Starting irqbalance ...
 * start-stop-daemon: did not create a valid pid in `/var/run/irqbalance.pid'                                                                         [ !! ]
 * ERROR: irqbalance failed to start

Code: Select all

depend_br0() {
rc_after_br0="net.eth0 net.vbox0"
}

net.eth0 and net.vbox0 won't exist after the upgrade (which is why those dependencies are failing). I imagine you'll need to create new symlinks named 'net.eth0' and 'net.vbox0', each with a target of 'net.lo'.

Check out the Gentoo handbook section on networking for more information on what you need to do.

Example:
cd /etc/init.d
ln -s net.lo net.eth0
rc-update add net.eth0 default

[energyman76b]

the migration guiide needs an update:
>It is critical that you run dispatch-conf and ensure your /etc is up to date before rebooting.

oh really? I never had that installed or planned to install it - and I am pretty sure that my system won't need it in the future. Or is dispatch.conf a base tool now?

[energyman76b]

and there seems to something wrong with the module-converter:
modules_2_6="${modules_2_6} r8169"
modules_2_6="${modules_2_6} ehci-hcd"

${modules_2_6}
?

[timeBandit]

the migration guiide needs an update:
>It is critical that you run dispatch-conf and ensure your /etc is up to date before rebooting.

oh really? I never had that installed or planned to install it - and I am pretty sure that my system won't need it in the future. Or is dispatch.conf a base tool now?

Oh really? Yes, for quite a while:

Code: Select all

gentoo@livecd ~ $ equery belongs $(which dispatch-conf)
[ Searching for file(s) /usr/sbin/dispatch-conf in *... ]
sys-apps/portage-2.1.2.2 (/usr/sbin/dispatch-conf -> ../lib/portage/bin/dispatch-conf)

Otherwise I agree--perhaps the guide should read:

It is critical that you run dispatch-conf or otherwise ensure your /etc is up to date before rebooting.

[energyman76b]

I am completly underwhelmed.

No, I am pissed.

fsck crashs out on boot - and then nothing goes, because it can not check some fs. WTF? ok, /fastboot works around it, but it is completly and absolutly fucked up.

what is wrong with this fstab - except that the old baselout worked flawlessly with it?
# /etc/fstab: static file system information.
# $Header: /var/cvsroot/gentoo-src/rc-scripts/etc/fstab,v 1.18.4.1 2005/01/31 23:05:14 vapier Exp $
#
# noatime turns off atimes for increased performance (atimes normally aren't
# needed; notail increases performance of ReiserFS (at the expense of storage
# efficiency). It's safe to drop the noatime options if you want and to
# switch between notail / tail freely.
#
# See the manpage fstab(5) for more information.

# <fs> <mountpoint> <type> <opts> <dump/pass>

# NOTE: If your BOOT partition is ReiserFS, add the notail option to opts.
/dev/sda1 /boot ext2 noauto,noatime 1 2
/dev/sda3 / reiserfs data=journal,noatime 0 1
/dev/sda2 none swap sw 0 0

#/dev/disk/by-uuid/7eda1c5e-c53d-4620-9a03-75c0a6e70774 /home reiser4 noatime,tmgr.atom_max_size=32768 0 2
/dev/disk/by-uuid/396a1f32-5f1b-42db-84f0-5ff54ca8ab6d /home reiser4 noatime,tmgr.atom_max_size=32768 0 2

/dev/sda6 /var reiser4 noatime,tmgr.atom_max_size=32768 0 2

/dev/disk/by-uuid/c1a84bd9-91f1-420e-9880-d4290599a399 /mnt/filme reiserfs data=journal,noatime,noauto,rw,users 0 0

/dev/disk/by-uuid/2a2f500a-125c-4205-b510-bd48e7b10c53 /opt reiser4 noatime,tmgr.atom_max_size=32768 0 0

# NOTE: The next line is critical for boot!

proc /proc proc nosuid,noexec 0 0

#proc /proc proc defaults 0 0


# glibc 2.2 and above expects tmpfs to be mounted at /dev/shm for
# POSIX shared memory (shm_open, shm_unlink).
# (tmpfs is a dynamically expandable/shrinkable ramdisk, and will
# use almost no memory if not populated with files)
shm /dev/shm tmpfs nodev,nosuid,noexec 0 0

#/tmp/swapfile none swap sw 0 0

/dev/cdrw1 /mnt/cdrom udf,iso9660 noauto,noatime,rw,users 0 0

tmpfs /var/tmp/portage tmpfs rw,size=2G 0 0
tmpfs /tmp tmpfs rw,size=512M,noexec 0 0

Second thing that REALLY pissed me of was the removal of the net stuff - ok. I was warned, but still.

Boot is not faster than before - no it is slower, same like reboot/shutdown.

And where can I tell the new baselayout/openrc stuff to NOT load modules except the ones explecitly told so?

oh and this:
fancontrol [ crashed ]

Love it! Not!