baselayout-2 issues

Message

gimpel · Post by **gimpel** » Wed Mar 26, 2008 11:56 pm

aldraia wrote:when the swap service has to start?

which program creates the xdm service? I've installed kdm,xdm, xorg and xorg-server and get nothing

thanks!

x11-apps/xinit

swap is started @ boot here.

n3Cre0 · Post by **n3Cre0** » Thu Mar 27, 2008 11:37 am

You need to add swap to bootlevel..

Code: Select all

rc-update add swap boot

About xdm: I don't know, I thought gdm or kdm.. maybe reinstall openrc?

UberLord · Post by **UberLord** » Thu Mar 27, 2008 1:14 pm

n3Cre0 wrote:About xdm: I don't know, I thought gdm or kdm.. maybe reinstall openrc?

Maybe read the post directly above yours

aldraia · Post by **aldraia** » Thu Mar 27, 2008 7:51 pm

Thank you. Everything now is working ok

red-wolf76 · Post by **red-wolf76** » Mon Mar 31, 2008 7:33 pm

Ok, it's bork your system time again!

Do I still need the -9999 ebuild or is the openrc-0.2-rc2 version currently in the portage tree sufficient? Do /etc/init.d, /etc/conf.d and bootlevels still require tweaking as per the last couple of posts?

How about an upgrade guide?

EDIT: Ok, emerged baselayout-2 and portage's openrc-0.2-rc2. updated all files marked by portage using dispatch-conf. Hopefully made all the adjustments to see that I retain all my settings. Especially, XSESSION seems to have gone missing. EDITOR went in /etc/profile.d/base.sh, so I'll try putting XSESSION in there as well.

Updated rc-config as per this thread:

Code: Select all

grei5003 ~ # rc-update show
                  xdm |      default
                 dbus |      default
                 famd |      default
                 fsck | boot
                 hald | boot
                 mtab | boot
                 ntpd |      default
                 root | boot
                 sshd |      default
                 swap | boot
              keymaps | boot
                acpid | boot
                boinc |      default
                cupsd |      default
                local |      default nonetwork
                samba |      default
           vixie-cron |      default
            syslog-ng |      default
           lm_sensors |      default
           localmount | boot
          consolefont |      default
              modules | boot
               esound |      default
               hdparm | boot
             hostname | boot
               net.lo | boot
             net.eth0 |      default
               procfs | boot
             netmount |      default
              urandom | boot
              numlock |      default
              hwclock | boot
            bluetooth |      default
             bootmisc | boot
        device-mapper | boot
            alsasound | boot

Unless bad things happened, I should be ready for rebooting, shouldn't I? Anyone? Bueller?

bblount · Post by **bblount** » Wed Apr 02, 2008 3:29 pm

Yeah I was having major problems with OpenRC until I read Roy's reply to add root,swap, etc.. to runlevel boot. Shouldn't that be an enotice when you install it?

Anyway, now things are working great.

gigs94 · Post by **gigs94** » Wed Apr 02, 2008 6:00 pm

UberLord wrote:For everyone having problems with the ebuilds in portage, please ensure the following

fsck, mtab, root, procfs are in the boot runlevel
checkroot and checkfs are NOT in the boot runlevel, and don't exist in /etc/init.d/
/bin/sh links to bash

Basically, the ebuild that entered portage isn't mine, and the Gentoo branch of OpenRC it pulls needs fixing a little which a Gentoo dev will have to do.

So this is *berry* *berry* important or you'll get a bunch of filesystem mounted read-only errors and will have to use some sort of rescue/live-cd to fix it. oops.

TuxTom · Post by **TuxTom** » Thu Apr 03, 2008 9:54 am

So is there any way to use dash instead of bash?

When I tried, udev didn't run as the kernel version check is buggy... And without udev no devices.

mv · Post by mv » Thu Apr 03, 2008 10:02 am

TuxTom wrote:So is there any way to use dash instead of bash?

With current openrc (also in the portage tree) this bug is fixed. However, better use the most current (~ARCH) version of dash - maybe earlier ones
don't have the required features.

gringo · Post by **gringo** » Thu Apr 03, 2008 11:18 am

anyone having troubles mounting /usr if it is a different partition ?
If i don´t manually hardcode /usr in the root script, the /usr partition doesn´t get even mounted at all.
This happens with every version i tried so far and with git as of yesterday - will open a bug later this week.

But thanks anyways for your work Roy, after a bit of testing everything apparently works as expected here

cheers

TuxTom · Post by **TuxTom** » Thu Apr 03, 2008 1:43 pm

mv wrote:However, better use the most current (~ARCH) version of dash - maybe earlier ones
don't have the required features.

Using app-shells/dash-0.5.4.6-r1 made it work. Thanks for the hint!

One thing is remaining: I get a message

Code: Select all

 * waitpid: Interrupted system call

right after

Code: Select all

INIT: version 2.86 booting

This doesn't seem to cause any further errors, but who knows for sure?

kernelOfTruth · Post by **kernelOfTruth** » Mon Apr 07, 2008 4:51 am

Posted: Tue Feb 12, 2008 3:07 am Post subject:

borh wrote:I have been having similar problems with the newest openrc/baselayout, which I tracked down to the rc_parallel="YES" flag. I tried getting a backtrace of the crash, but it seems to only trigger when I am not in maintenance (console) mode. I could start all the services by hand and get to X in maintenance mode, which led me to suspect that the parallel loading was somehow messing up. Getting out of the console back to the interactive mode, the system would always stop at "Caching service dependencies" (which I'm guessing is related to device-mapper). A Ctrl-C would only print out something like what has already been posted here, but since the system was dead (sometimes a hard reboot was needed) I couldn't record the error. I'll try tomorrow to see if there's any way I can duplicate the error inside gdb.

this also happens to me on my laptop - NOW !

it caches on almost every step, between some services also twice or more often

the system is usable after a long time I have dm-crypt loaded at startup with it but everything commend out in /etc/conf.d/dmcrypt (in contrast to my desktop, might this be the culprit ? --> I'll have a look at that later today) ...

this also happens during shutdown and/or reboot - so something is really messed up

I copied over my desktop system to laptop via stage4-method (/etc/hosts and /etc/conf.d/hostname, /etc/conf/dmcrypt changed

but how to fix this ?

UberLord · Post by **UberLord** » Mon Apr 07, 2008 7:23 am

gringo wrote:anyone having troubles mounting /usr if it is a different partition ?
If i don´t manually hardcode /usr in the root script, the /usr partition doesn´t get even mounted at all.
This happens with every version i tried so far and with git as of yesterday - will open a bug later this week.

As stated earlier, please ensure that mtab, fsck and root have been added to your boot runlevel.

UberLord · Post by **UberLord** » Mon Apr 07, 2008 7:26 am

If services are constantly caching depends, it means that a file in /etc has an mtime in the future, which is bad.
The normal cause of this is to have /etc/localtime as a symlink instead of a real file, and mainly seems to affect people in >=+1 GMT timezones.

gringo · Post by **gringo** » Mon Apr 07, 2008 7:36 am

As stated earlier, please ensure that mtab, fsck and root have been added to your boot runlevel.

yes, they are of course, root mounts correctly but /usr is completely ignored here in a default setup.
Havnt´had much time and i have limited access to this lappy, but will fill a bug asap.

cheers

kernelOfTruth · Post by **kernelOfTruth** » Mon Apr 07, 2008 7:55 am

UberLord wrote:If services are constantly caching depends, it means that a file in /etc has an mtime in the future, which is bad.
The normal cause of this is to have /etc/localtime as a symlink instead of a real file, and mainly seems to affect people in >=+1 GMT timezones.

thanks for that info,

(un)fortunately it's a file:

ls -l /etc/localtime
-rw-r--r-- 1 root root 2211 17. Mär 00:23 /etc/localtime

UberLord · Post by **UberLord** » Mon Apr 07, 2008 8:21 am

Does it re-cache each time you restart a service?
If so, run this command

find /etc -mtime -0
And that will (should) list all files with future mtimes that could be affecting the issue.

kernelOfTruth · Post by **kernelOfTruth** » Mon Apr 07, 2008 8:37 am

UberLord wrote:Does it re-cache each time you restart a service?
If so, run this command

find /etc -mtime -0
And that will (should) list all files with future mtimes that could be affecting the issue.

I just started up the laptop & the problem "fixed" itself,

the problem must have been caused by the livecd & used to extract the stage4 tarball (afaik the time was 9 hours ahead, now at least 9 hours have been passed)

I should get accustomed to set the timezone & time properly when using a livecd

thanks again !

Thargor · Post by **Thargor** » Mon Apr 07, 2008 11:45 am

UberLord wrote:The normal cause of this is to have /etc/localtime as a symlink instead of a real file, and mainly seems to affect people in >=+1 GMT timezones.

So it should be a real file? I am asking because I thought that it should only be a file if people have /usr on a separate partition, what would lead to /etc/localtime not being available during boot.

UberLord · Post by **UberLord** » Mon Apr 07, 2008 12:16 pm

Thargor wrote:
UberLord wrote:The normal cause of this is to have /etc/localtime as a symlink instead of a real file, and mainly seems to affect people in >=+1 GMT timezones.
So it should be a real file? I am asking because I thought that it should only be a file if people have /usr on a separate partition, what would lead to /etc/localtime not being available during boot.

If /usr is on / then localtime can be a symlink

n3Cre0 · Post by **n3Cre0** » Sat Apr 12, 2008 10:41 am

Hey

I was wondering if splashutils is working for someone now?

I noticed

# ls -al /usr/portage/media-gfx/splashutils/files/
total 40K
drwxr-xr-x 2 paludisbuild paludisbuild 4.0K Apr 11 18:06 ./
drwxr-xr-x 3 paludisbuild paludisbuild 4.0K Apr 11 18:06 ../
-rw-r--r-- 1 paludisbuild paludisbuild 292 Apr 11 18:01 openrc-splash-crash.patch
-rw-r--r-- 1 paludisbuild paludisbuild 9.7K Mar 25 10:00 openrc-strlist-abi.patch
-rw-r--r-- 1 paludisbuild paludisbuild 15K Oct 11 2007 splashutils-1.5.2.1-baselayout-rc5.patch

But a recompile still gives me

# /etc/init.d/fbcondecor restart
* Setting framebuffer console images ...
/sbin/splash-functions.sh: line 481: 4766 Segmentation fault ${spl_decor} --tty="${tty}" -t "${theme}" -c setcfg
* Failed to set background image on tty1 [ !! ]
* ERROR: fbcondecor failed to start

octoploid · Post by **octoploid** » Mon Apr 14, 2008 3:51 pm

OK I migrated today and everything is working again.

But there were two issues that I thought I might mention:

- The network configuration syntax changed.
I had to change my original config:

Code: Select all

config_eth1=( "192.168.0.2 netmask 255.255.255.0 broadcast 192.168.0.255 mtu 1492" )
routes_eth1=( "default gw 192.168.0.1" )

to:

Code: Select all

config_eth1=( "192.168.0.2 netmask 255.255.255.0 broadcast 192.168.0.255" )
routes_eth1=( "default via 192.168.0.1" )                                                                                                          
mtu_eth1="1492"

IMHO this should be mentioned in the Migration Guide.

- My dmcrypt partition did not work initially. I had to deinstall cryptsetup-luks.
After that I emerged cryptsetup and moved /etc/conf.d/cryptfs to /etc/conf.d/dmcrypt.

Jupiter1TX · Post by **Jupiter1TX** » Mon Apr 14, 2008 7:51 pm

n3Cre0 wrote:Hey

I was wondering if splashutils is working for someone now?

Strange thing my splash works sometimes and not others.
I see no errors anywhere but it only works sometimes.

UberLord · Post by **UberLord** » Mon Apr 14, 2008 8:12 pm

Jupiter1TX wrote:Strange thing my splash works sometimes and not others.

Two crashers for splash have been identified and fixed in the git repos for both. OpenRC-0.2.2 or newer will have the fix. I think the latest splashutils also has a related fix in portage but has not been bumped (and I've not verified).

no4b · Post by **no4b** » Tue Apr 15, 2008 12:38 pm

Once I upgraded to baselayout-2 /etc/conf.d/net.eth0 and /etc/init.d/net.eth0 disappeared. I recreated /etc/conf.d/net.eth0 but I don't know where I can find example of /etc/init.d/net.eth0.