I installed gentoo system from scratch. When I try to boot gentoo system first time I get error message saying: Dependency info missing! Please run # /sbin/depscan.sh to fix this. That leads me to file /lib/rcscripts/sh/rc-services.sh where this error is thrown if running of [ "${RC_GOT_DEPTREE_INFO}" != "yes" -a -f "${svcdir}/deptree" ] \ && source "${svcdir}/deptree" is failed. ${svcdir} is defined in /etc/conf.d/rc and it it /var/lib/init.d by default. Looking at scripts in /etc/init.d/ tells me that deptree is required by modules. When I look at ${svcdir}/deptree I noticed that /etc/init.d/localmount (mounting /var and other partitions) is done after /etc/init.d/modules. This means that partitions are not mounted at the time ${svcdir} is accessed and this causes the error message to appear. Reproducible: Always Steps to Reproduce: Basic installation with separated /var partition. Actual Results: At boot error specified in details part is thrown after
I installed gentoo system from scratch. When I try to boot gentoo system first time I get error message saying: Dependency info missing! Please run # /sbin/depscan.sh to fix this. That leads me to file /lib/rcscripts/sh/rc-services.sh where this error is thrown if running of [ "${RC_GOT_DEPTREE_INFO}" != "yes" -a -f "${svcdir}/deptree" ] \ && source "${svcdir}/deptree" is failed. ${svcdir} is defined in /etc/conf.d/rc and it it /var/lib/init.d by default. Looking at scripts in /etc/init.d/ tells me that deptree is required by modules. When I look at ${svcdir}/deptree I noticed that /etc/init.d/localmount (mounting /var and other partitions) is done after /etc/init.d/modules. This means that partitions are not mounted at the time ${svcdir} is accessed and this causes the error message to appear. Reproducible: Always Steps to Reproduce: Basic installation with separated /var partition. Actual Results: At boot error specified in details part is thrown after ´checkroot´ and ´hostname´ scripts are ran. At ´calculating module dependencies´ comes the error which causes automatic shutdown of the gentoo. Expected Results: No automatic shutdown of gentoo.. =) /etc/init.d/localmount is run before /etc/init.d/modules. baselayout-1.8.6.10-r1 and -1.8.6.12-r3 tried portage 2.0.49-r21 used
Basically you need to point your svc_dir to some location on the root partition
Changing svcdir to /lib/boot/init.d in /etc/conf.d/rc but that isn't helping. Getting still the same error. This all worked like a dream before. I have used gentoo 1 * years without this kind of problems...
The boot works as follow: - Run the following services in listed order by hand (_before_ running depscan.sh, so it do not matter if /var is not mounted): checkroot hostname modules checkfs localmount - Run depscan.sh - Run the rest of the services at boot Meaning, if your /var is in /etc/fstab, and no issues, it should be mounted _before_ /var/lib/init.d/ is accessed ... Does /var/lib/init.d (on the actual /var), /lib/boot/init.d or whatever you actually now use as $svcdir exist?
You are right about that. I checked out the boot again and find out that it is 'Calculating module dependencies' which crashes and shuts down the system. It doesn't give any error. /sbin/depscan.h error message shows up after this but it seems that it thrown by shutdown procedures. I still don't know what is wrong about my system. I have installed the system twice and both times I have had the same kind of error situation. Perhaps I should try to install my gentoo again from the scratch and hope the best because I think modules cannot be removed from the boot because of the dependencies. I don't have any kernel modules defined in /etc/modules.autoload/kernel-2.6.
So this is only on shutdown? Any possible way to get a log through serial console or such ?
Yes, this /sbin/depscan.sh error comes only in shutdown after modules-script is crashed. I suppose that at this stage there is no logger to log anything and in screen there is nothing special to see. ... Remounting root filesystem read/write... [ok] Setting hostname to main... [ok] Calculating module dependencies... INIT: Switching to runlevel: 0 INIT: Sending processes the TERM signal ...
Yes, /sbin/depscan.sh error is coming only after modules-script is crashed. I suppose system loggers are not working at this point and there is nothing special to see in the screen. ... Remounting root filesystem read/write... [ok] Setting hostname to main... [ok] Calculating module dependencies... [ok] INIT: Switching to runlevel: 0 INIT: Sending processes the TERM signal ...
I tried to debug the cause of the crash and the traces lead to /sbin/modules-update and to the following lines: ... export TESTING_PROBE_CONF="${TMPFILE3}" if /sbin/generate-modprobe.conf >> "${TMPFILE4}" 2> /dev/null then ... That means /sbin/generate-modprobe.conf have to be crashing somehow. I find this out by echoing numbers thru scripts since I get into this point. I am not sure how to debug this further.
is this still a problem?
Not anymore. I got the system running couple of days after last messages and haven't faced the same problem after that.
everything seems to be sane now with baselayout 1.11.x
