I've just emerged the new baselayout (version 1.8.6.2) as usual, i've upgraded all 27 files with etc-update, updated env vars with env-update ;p... The day after i turn on my pc and i can't boot... Seems strange but i get a bunch of errors like: Feb 5 17:36:35 tiger rc-scripts: ERROR: "/var/state/init.d/started/checkfs" has syntax errors in it; not executing... Feb 5 17:36:35 tiger rc-scripts: ERROR: "/var/state/init.d/need/checkroot/checkfs" has syntax errors in it; not executing... These are the only two that my syslog registered but there are pages and pages of these errors (also complaining about the non-existance of softscript.new and softscript.old... but during the emerge i've got no errors)... Have i done some mistakes of is there a problem with this new package? I've updated baselayout also on another machine but i think that i won't reboot for now ;p... Thanx a lot Que Reproducible: Always Steps to Reproduce: 1. emerge -u baselayout 2. etc-update 3. reboot 4. urgh :)) Actual Results: after all it crashes :)
I had the problem where when I rebooted, I got the missing link from /var/state to /mnt/.init.d, because I had not done the etc-update (emerged overnite and forgot). not it boots fine with just a couple minor flaws.. 1) starting devfsd gives output, which it never did before(e.g. starting devfsd management daemon 1.3.x blah blah..) my guess is it isnt pointing at /dev/null 2)I know devfsd is supposedly started and ps shows it, but now all my filesystem output is showing the newstyle names even though devfsd is running and the symlinks are there. when I do mount I see stuff like : rootfs on / type rootfs (rw) /dev/root on / type ext3 (rw) and so on when I should see for those two entries: /dev/hdb2 on / type ext3 (rw) the other curious thing is since the move to /var/state/init.d I dont have tmpfs mounted on /mnt/.init.d anymore...will explore that to see if it helps.
when I do mount I see stuff like : rootfs on / type rootfs (rw) /dev/root on / type ext3 (rw) and so on when I should see for those two entries: /dev/hdb2 on / type ext3 (rw) ---------------------------------- I decided to just cp /proc/mounts to /etc/mtab, as its less prone to fail than the "try to recreate all /etc/mtab entries with fake mounts". It does not do any harm, but if it really bothers that much ...
Well, I realize that its probably more foolproof that way, and I realize that I may be alone in this, but I'm not a big fan of devfs on the desktop, so I like to see the old names. Plus if it was just that one partition I could probably get over it,but I have 6 partitions on my Gentoo install, and mount another 5 on the other disk and /dev/ide/host0/bus0/target*/lun0/part* 15 times looks crazy when doing mount or df. Also, could you point me at where I can > /dev/null the starting devfsd output in init? I can see the vesion in dmesg when it rolls by and the output spoils the pretty green messages:)
Uhm, i've just noticed that my reiserfs is mounted readonly with this version of baselayout... Anyway fstab is ok... Dunno why. Que
I just emerged the this baselayout and rebooted and I get a ton of messeges on the console inbetween initscripts like: modprobe: Can't locate module /dev/parallel modprobe: Can't locate module /dev/parallel modprobe: Can't locate module /dev/lp1 modprobe: Can't locate module /dev/par1 modprobe: Can't locate module /dev/printers/1 modprobe: Can't locate module /dev/parallel modprobe: Can't locate module /dev/lp2 modprobe: Can't locate module /dev/par2 modprobe: Can't locate module /dev/printers/2 modprobe: Can't locate module /dev/parallel modprobe: Can't locate module /dev/lp3 modprobe: Can't locate module /dev/par3 modprobe: Can't locate module /dev/printers/3 modprobe: Can't locate module /dev/ttyS2 modprobe: Can't locate module /dev/ttyS3 modprobe: Can't locate module /dev/ttyS4 modprobe: Can't locate module /dev/ttyS5 modprobe: Can't locate module /dev/ttyS6 modprobe: Can't locate module /dev/ttyS7 modprobe: Can't locate module /dev/ttyS8 modprobe: Can't locate module /dev/ttyS9 modprobe: Can't locate module /dev/ttyS10 modprobe: Can't locate module /dev/ttyS11 modprobe: Can't locate module /dev/ttyS12 modprobe: Can't locate module /dev/ttyS13 modprobe: Can't locate module /dev/ttyS14 theres about 3-4 times that many but thats all I grabed before it went by and it doesnt show up in dmesg or the syslog.
and the output of df is really ugly...what happend to the old setup? Filesystem Size Used Avail Use% Mounted on /dev/root 14G 12G 1.5G 89% / /dev/ide/host0/bus0/target1/lun0/part2 9.2G 8.7G 13M 100% /mnt/linuxback /dev/ide/host0/bus0/target1/lun0/part1 4.7G 4.7G 40M 100% /mnt/winback /dev/ataraid/disc0/part1 86G 54G 33G 62% /mnt/raid /dev/ide/host0/bus1/target1/lun0/part1 75G 74G 634M 100% /mnt/anime /dev/ide/host2/bus1/target0/lun0/part1 112G 109G 3.2G 98% /mnt/stuff /dev/ide/host2/bus1/target1/lun0/part1 75G 74G 1.3G 99% /mnt/tv I know previous comments mentioned this but I didnt know it was this bad until I ran it.
Oh and another thing...after emerging this baselayout and doing etc-update, mozilla stoped working. I run it and it doesnt do any disk access and it uses 95% cpu and a ever increasing amount of ram until I kill it. I searched for mozilla but all the bugs were about crashes or compile problems. I emerged a few other packages before baselayout, but there not anything that I could see effecting mozilla. I tried remerging mozilla to no effect. Galeon runs but keeps opening more and more windows and doesnt load any pages. I'm stuck using pheonix-bin which renders ugly compared to mozilla.
and apperently pheonix doesn't do line wraps when posting the comments. sorry about that.
This one is biting a lot of people. Perhaps baselayout should display a big red 'YOU MUST RUN etc-update' notice and kill portage so people notice it?
I think that pkg_postinfo should have to be displayed after the clean phase, btw i've done etc-update as usual but now my fs is mounted readonly, the only thing i can do is to pass "linux rw" to lilo, then i can enter in management-mode but still i can't boot and i get a bunch of errors in the proc fs... What should i do? :) thanx a lot again :)
My opinion regarding comment #9: YES!!!!! I haven't run into any problems with the new baselayout (4 machines) so far - but, having been burned once before I take the following steps: # emerge rsync # emerge -pu world - notice new baselayout update - BY ITSELF!!! # emerge baselayout # etc-update This works, provided you make good choices during etc-update. The phrase 'good choices' is critical to a successful result. I've had as many as 58 config files to be updated. And it's easy to make a mistake. Also, during 'etc-update' if you're thinking of using option 2 (keep current config and abandon new config), you should regard this as a mistake. You should be using either options 1 or 3 only for each changed config file. I've noticed that some of the changes are very subtle but important. After (and ONLY after) the emerge of baselayout and running of etc-update are complete do I continue with the rest of the emerges.
I also did etc-update and I took all the new configs except one, and I mereged the changes of that one, so I doubt my problems are becaue of not doing or doing etc-update imporperly. I also merged baselayout by itself and ran etc-update imideatly after emerge.
I get ReadOnly system, can't create symbolic linca bla bla...and hangs 500 lines like this /sbin/rc: line 275: /var/state/init.d/softlevel: Read-only file system bla bla
The same here, azarah is this an ebuild problem or we are doing some kind of mistakes? thanx a lot Que
There also seems to be an problem with the /etc/init.d/net.eth0 script for some people after upgrading to this baselayout version. Having net.eth0 added to the default run level causes * Bringing eth0 up... [ ok ] * Setting default gateway... SIOCADDRT: File exists * Failed to bring eth0 up [ !! ] for some of us as discussed at http://forums.gentoo.org/viewtopic.php?t=33962 Removing net.eth0 from the default runlevel seems to fix this issue, and eth0 seems to be brought up all on its own.
I agree with Comment #9 I usually don't update my inittab file because I change the tty to vc/ as I don't use compatability mode with devfs. Because I didn't update inittab, my system was broken after upgrading to baselayout-1.8.6.2. I question if every Gentoo always updates thier etc files after an upgrade. Most of the time not updating etc files isn't a big deal.
The issue with ugly `df` output is unrelated to baselayout. This is from util-linux - in z, util-linux does #define resolve_symlinks at the beginning of mount/realpath.c - in y, util-linux undefs it. Easy enough to fix...
Just committed util-linux-2.11z-r1, which should fix the ugly df output.
I dont suppose util-linux is gifted enough to fix the duplicate rootfs entries and point {mount|df} at the compat symlinks? This would work exceedingly well with azarah's choice to cp /proc/mounts to /etc/mtab. rootfs on / type rootfs \____/dev/hdb2 on / type ext3 /dev/root on / type ext3 / Heres to hoping!BTW nice job avenj!
Seth, sorry to pin this on you, but you unmasked baselayout-1.8.6.2 by mistake :( It was still very experimental, as users usually do not etc-update, or read post messages :/ http://cvs.gentoo.org/cgi-bin/viewcvs.cgi/gentoo-x86/profiles/package.mask.diff?r1=1.2007&r2=1.2008
yeah i'll have to take credit for this problem you want to remask it? i think we probably need to automate some of the processes before we can unmask it again... seth
If you remask it, the same thing happens again (in terms of people not running etc-update and having an unbootable system) - but it's probably a good idea anyway...
*** Bug 15281 has been marked as a duplicate of this bug. ***
*** Bug 15185 has been marked as a duplicate of this bug. ***
baselayout-1.8.6.4 has been out a while, and all issues addressed.