openrc support thread

[slackline]

@slack---line:
Have you tried to change the "strict" variable I mentioned some posts ago ? This one helped me with the similar problem of net.eth0 / net.eth1.

Cheers for that drizzt, I'd missed your posts and like yourself not read the comments in /etc/rc.conf in sufficient detail.

[zAfi]

just upgrade to b2/openrc and it works like a charm!
There's just one thing, I get "err, no interface specified" messages all the time I start/stop/restart a net.* service, although everything works.

Code: Select all

~ $ sudo /etc/init.d/net.wlan0 restart
 * Caching service dependencies ...
err, no interface specified
err, no interface specified
err, no interface specified                                                                                                                           [ ok ]
err, no interface specified
err, no interface specified
 * Bringing down interface wlan0
 *   Stopping dhcpcd on wlan0 ...                                                                                                                     [ ok ]
 *   Stopping wpa_cli on wlan0 ...                                                                                                                    [ ok ]
 *   Stopping wpa_supplicant on wlan0 ...                                                                                                             [ ok ]
 *   Running postdown ...
err, no interface specified
 * Bringing up interface wlan0
 *   Running preup ...
 *   Starting wpa_supplicant on wlan0 ...
ioctl[SIOCSIWAUTH]: Operation not supported
WEXT auth param 4 value 0x0 - ioctl[SIOCSIWAUTH]: Operation not supported                                                                             [ ok ]
 *   Starting wpa_cli on wlan0 ...                                                                                                                    [ ok ]
 *   Backgrounding ... ...

It seems to me that dhcpcd tries to do something before it actually invokes the device, but this is only a weird idea.

Here's /etc/conf.d/net, it should be alright according to the upgrade guide.

Code: Select all

config_eth0="dhcp"
dhcpcd_eth0="-t 5" #Timeout after 10 seconds
dhcp_eth0="release nontp nonis" #Only get an adress

config_wlan0="dhcp"
dhcpcd_wlan0="-t 5"
dhcp_wlan0="release nontp nonis"
# Prefer wpa_supplicant over wireless-tools
modules="wpa_supplicant" "dhcpcd"
wpa_supplicant_wlan0="-Dwext -B -qq -c/etc/wpa_supplicant/wpa_supplicant.conf"

#preup() and postdown() are needed to un-/load the wlan module!
#By simply restarting the service without reloading the module
#we won't receive an IP anymore.
preup() {
  if [ $IFACE == "wlan0" ]; then
    modprobe iwl4965 && sleep 0.5 #wait 1 second before continuing, else we fail!
  fi
}
postdown() {
  if [ $IFACE == "wlan0" ]; then
    modprobe -r iwl4965
  fi
}

I already reemerged dhcpcd and openresolv but to no avail. Any ideas

[sdtester]

The only problem I experienced with upgrading to b2/openrc is my dummy net interface is now not doing its job. I have the dummy interface there so that apache, named and other services will run when I don't have another internet connection. dummy0 is getting coming up with an ip address (10.0.0.2) but the other services remain inactive:

# rc-status
Runlevel: default
...
net.MainLAN [ inactive ]
net.dummy0 [ started ]
named [ scheduled ]
...
netmount [ scheduled ]
...
sshd [ scheduled ]
apache2 [ scheduled ]
slapd [ scheduled ]
samba [ scheduled ]
...

dummy is compiled into the kernel as a module.
When I execute "/etc/init.d/net.dummy0 restart" this is spewed out:

* Bringing down interface dummy0
* Stopping netplug on dummy0 ... [ ok ]
* Bringing up interface dummy0
* Starting netplug on dummy0 ... [ ok ]
* Backgrounding ...
* WARNING: net.dummy0 not under our control, aborting

[UberLord]

Do you have iproute2 installed?

[sdtester]

Yup, iproute is current at 2.6.24.20080108. I even made it explicit in /etc/conf.d/net (modules="iproute2")

dummy0 is not even getting a mention as being able to satisfy the 'net' dependency for services like apache. For example,

Code: Select all

# /etc/init.d/apache2 restart
 * WARNING: apache2 is scheduled to start when net.MainLAN has started

It use to be dependent on either MainLAN or dummy0 when using baselayout-1

[UberLord]

It's because they're both in the runlevel, so both are needed to satisfy the dep.
If only net.dummy0 is meant to satisfy the dep then remove net.MainLAN from the default runlevel or remove it's net provision like so in /etc/conf.d/net

Code: Select all

rc_net_MainLAN_provide="!net"

[slackline]

Or you could perhaps set...

Code: Select all

# Do we allow any started service in the runlevel to satisfy the depedency
# or do we want all of them regardless of state? For example, if net.eth0
# and net.eth1 are in the default runlevel then with rc_depend_strict="NO"
# both will be started, but services that depend on 'net' will work if either
# one comes up. With rc_depend_strict="YES" we would require them both to
# come up.
rc_depend_strict="NO"

...as I was advised further up the thread.

[sdtester]

Thanks Uberlord, That did the trick!

[elcugo]

Hello, I just upgraded to b2/openrc. My system seems to boot faster, saves me 23 seconds, yay!

I have problems with my net config. I had:

Code: Select all

modules=( "iproute2" )
modules=( "dhcpcd" )
config_eth0=( "dhcp" )
fallback_eth0=( "192.168.1.66/24" )
dhcpcd_eth0=( "-t 25" )

I tried just removing the "()"s but I get errors in the console. How do I migrate that config?

This is what I have right now:

Code: Select all

config_eth0= "dhcp"
fallback_eth0= "192.168.1.66/24"

But it complains about those commands not found.

Fortunatelly the script falls back to using dhcpcd anyway so I get my NIC configured.

[golding]

Try this, has been working for me after I upgraded to bl-2 and openrc

Code: Select all

modules="iproute2"
config_eth0="dhcp 192.168.99.1/24" 
dhcp_eth0="nontp"

An example file can be found at /usr/share/doc/openrc/net.example

[elcugo]

Never mind, I just realized that I forgot to remove whitespace after the "=". Foolish me.

An example file can be found at /usr/share/doc/openrc/net.example

Ahh. I was looking at /etc/conf.d/net.example, which is useless. Thanks! Now my config works:

Code: Select all

config_eth0="dhcp"
fallback_eth0="192.168.10.100/24"
dhcpcd_eth0="-t 25"

No need to add modules="iproute2" and modules="dhcpcd" since they seem to be the defaults now.

[crowbert]

I finally got ifplugd working.

I had to upgrade dhcpcd to the latest version.
I had to add -L to dhcpcd options to prevent it from configuring disconnected interfaces.
I had to set rc_depend_strict="NO" in rc.conf so that services would start without both interfaces coming up.

I also found that ifplugd works correctly with parallel initialization. My previous experience was that ifplugd and parallel startup was not a happy configuration. (I thought there was a bug report on this, but I couldn't find one.)

Anyway, much of my difficulty was due to errors in reconstructing /etc/conf.d/net. The rest was in not paying enough attention to /etc/rc.conf.

[semdornus]

Just a small fyi; I'm using dash as my /bin/sh and I got the message "/lib/rc/sh/runscript.sh: 193: [[ not found"
when starting cpufreqd. Of course after removing all the double brackets it works as expected.

[fintan]

Hi all

I've just installed openrc. I had some issues but most of them are sorted now. My home partition is 200G+. So it takes some time to check. I was wondering is it possible to check it in the background since it's not needed until I login? Thanks

fintan

[UberLord]

I've just installed openrc. I had some issues but most of them are sorted now. My home partition is 200G+. So it takes some time to check. I was wondering is it possible to check it in the background since it's not needed until I login? Thanks

Check /etc/conf.d/fsck for an example of this.
You should enable rc_parallel in /etc/rc.conf for best benefit though.

[fintan]

Thanks UberLord

So if I set fsck_passno="=1" to one it should just check one file system a a time?

[UberLord]

No, it will just check fstab entries marked with a passno of 1.

[fintan]

cool I'll figure the rest out. Thanks again

[furanku]

Here my experience from the migration:

- Besides that also i needed to recreate net.eth0,
- /etc/init.d/kexec stopped working (rebootet normally via bios, not kexec). I replaced it with the script from bugs.gentoo.org, which made it working again, but autodetection of kernel commandline and root device still didn't work. Entering these in /etc/conf.d/kexec was a workaround for that.
- In /etc/init.d/vboxdrv I had to replace /sbin/functions.sh with /etc/init.d/functions.sh. Is there already a bug report for that?
- I have a non-gentoo initscript for the commercial twonkyserver (UPnP) which stopped working if started in the boot process, but works if called manually via "/etc/init.d/twonkyserver start". Have to investigate that further, help would be appreciated
- Due to the almost same time upgrade to 2.6.25 and the problems with uvesafb, I lost my framebuffer console, but I'm not sure if BL2/OpenRC is involved into that and repaired it by setting it up from scratch.

Boot time is greatly improved, all in all it was worth it although the transisition was a little bit rough due to the buggy first 2.6.25 release, should have known better not to change two imporant parts of the system at the same time.

[NiceGuy]

Hello,

Other than a faster bootup time, can I get some clarity on what other benefits there are in adopting openrc on my system? I have just recently switched, and I'll admit, it appears nice, quick and clean, but I'm new to it, and am not entirely sure if I am utilizing it to its maximum.

Are there other current benefits or perhaps just long terms benefits that I am not aware of?
Thanks

[darkphader]

Still have a number of issues with openrc.

• alsa settings don't save and restore
• nsd doesn't start, which means that:
  • named doesn't start
  • mit-krb5kdc doesn't start
  • mit-krb5kadmind doesn't start
• txquelen doesn't get set (oddly enough mtu does)

In my case bind is just the dns cache, nsd is the authoritative server, and kerberos needs the TXT and SRV records in order to start. It all worked fine before the "upgrade". The alsa problems are with my desktop (it's not running on the server).

Chris

[bunder]

merged one post.

[UberLord]

You should open bugs for those as I can't find any existing bugs.

[darkphader]

You should open bugs for those as I can't find any existing bugs.

The txqueuelen bug was opened on 4/16:
http://bugs.gentoo.org/show_bug.cgi?id=217948

Will get to the others - waiting to be at the console when nsd fails so I can provide more info.

[trad511]

...You need to get dmcrypt to run after root. Try putting this in /etc/conf.d/dmcrypt

rc_need="root"

rc_need="root" and rc_dmcrypt_need="root" both result in dmcrypt failing to start since "fsck would fail". I also tried removing fsck, rc_dmcrypt_after="root", etc..., but to no avail. In addition the depends() functions in root, fsck and dmcrypt seem to force the requirements I need, but still doesn't work.

I have a gpg key on usb to decrypt /home, but the /lib/rcscripts/addons/dm-crypt-start.sh requires / to be rw as shown below:

Code: Select all

local mntrem=/mnt/remdev.$$
if [ ! -d "${mntrem}" ] ; then
    if ! mkdir -p "${mntrem}" ; then
        ewarn "${source} will not be decrypted ..."
        einfo "Reason: Unable to create temporary mount point '${mntrem}'"
        return
    fi
fi

Unfortunately it still seems like fsck, root and dmcrypt create a circular dependency for startup (despite trying to override) which dmcrypt always loses.

Any other ideas?

I've solved this.


I have /home, /tmp, and swap in dmcrypt, but this would probably work for everything except encrypted /.

The suggestion I was given by Roy was to put rc_need="root" in /etc/conf.d/dmcrypt. Unfortunately, as stated in my last post dmcrypt needs /mnt to be read-write in order to temporarily mount the usb drive to get the key (in my case gpg-encrypted). The real problem is that rc_need="root" puts root, fsck and dmcrypt in a circular dependency which dmcrypt always lost.

But, using Roy's suggestion I tried several options. I needed:

1. dmcrypt needs / to be read-write
2. dmcrypt needs to run after fsck (bear with me on this one) to break the dependency cycle.
3. dmcrypt needs to run before localmount or else filesystems (in fstab) won't get mounted.
4. localmount mounts things in order of appearance in /etc/fstab.

There are probably some simplifications that can be made (will try them soon), but this is what I got to work:

1. Add to /etc/rc.conf: rc_dmcrypt_need="root"
2. Add to /etc/rc.conf: rc_dmcrypt_after="fsck"
3. Add to /etc/rc.conf: rc_dmcrypt_before="localmount"
4. Put encrypted filesystems last in /etc/fstab (maybe not needed)
5. Put sixth field (fsck) in /etc/fstab to zero (0) for all encrypted filesystem (maybe not needed).

This works on both my systems and does not require script hackery - it uses the super flexible native facilities built in to openrc. The only thing I don't like about the solution is the mods to /etc/fstab. This shouldn't be needed (and may not, but I've not tested).

I believe that a bit better HOWTO of openrc rc_ options and boot ordering and baselayout2 would have sped up my discovery, but a week of (un)concentrated work isn't bad for a brand new application. I do think, after looking through cryptsetup scripts that _assuming_ /mnt will be read-write at the time the init script is run is not great design. A key exchange on a ramdisk or temporary ramdisk mount point might be better.

Hope this helps somebody. I'll let you know when I try the other (non-fstab hack) options.