small gotcha, if you want to mount nfs filesystems at boot time the sensible thing to do is `rc-update add portmap default` (and netmount if it's not already) The trick is, with just portmap and netmount added, netmount runs before portmap and remote FS's fail to mount, if you `rc-update add nfs default` as well, it makes portmap start first and consequently netmount runs properly. should we make netmount need portmap? right now, I've just left nfs enabled but I'd like to turn it off again sometime.
The bug which causes the start og netmount before the portmapper lies in line 24 of /etc/init.d/netmount: 'need ${mydeps}'. But the dependecies for the netmount service is calculated in the variabel myneed. So line 24 should look like this: 'need ${myneed}' if i'm not wrong. BTW i'm talking about netmount 1.17 2002/08/25 20:29:07 from baselayout 1.8.3
Thanks, small type-o from last time i was on crack :-) Anyhow, fixed on cvs.
*** Bug 7912 has been marked as a duplicate of this bug. ***
*** Bug 7992 has been marked as a duplicate of this bug. ***
*** Bug 8226 has been marked as a duplicate of this bug. ***
*** Bug 8261 has been marked as a duplicate of this bug. ***
could you double check this ? i ran `ebuild baselayout-1.8.3.ebuild install` right after `cvs update` and the resulting ${D}/etc/init.d/netmount file still has this bug ...
Fixed on CVS. I was still waiting for the damn raid thing to get resolved.
*** Bug 8412 has been marked as a duplicate of this bug. ***
*** Bug 8559 has been marked as a duplicate of this bug. ***
*** Bug 8670 has been marked as a duplicate of this bug. ***
bug is still present. i just ran into it.
Attach your copy of netmount.
Created attachment 4887 [details, diff] One line patch correcting typo causing portmap to start before netmount I just built this box and the bug still seems to be in the netmount script as of 10/19/2002. Here's a one line patch that fixes it.
its been fixed in all my systems with 1.8.4 ... root@vapier eggdrop # cat /etc/gentoo-release Gentoo Base System version 1.4.1 root@vapier eggdrop # emerge -s baselayout Searching... [ Results for search key : baselayout ] [ Applications found : 1 ] * sys-apps/baselayout Latest version available: 1.8.4 Latest version installed: 1.8.4 Homepage: http://www.gentoo.org/ Description: Base layout for Gentoo Linux filesystem (incl. initscripts and sysvinit) root@vapier eggdrop # ebuild /usr/portage/sys-apps/baselayout/baselayout-1.8.4.ebuild install >>> md5 ;-) sysvinit-2.84.tar.gz >>> Unpacking source... >>> Unpacking sysvinit-2.84.tar.gz >>> Unpacking rc-scripts-1.4.1.tar.bz2 >>> Source unpacked. [ lots of stuff ] >>> Completed installing into /var/tmp/portage/baselayout-1.8.4/image/ root@vapier eggdrop # grep need /var/tmp/portage/baselayout-1.8.4/image/etc/init.d/netmount local myneed="net" local myneed="${myneed} portmap" need ${myneed} root@vapier eggdrop #
The problem is: i don't know where do you get the 1.8.4 from. I just made an 'emerge rsync' on my 1.2 box and it looks like this: cloudcity root # emerge -s baselayout Searching... [ Results for search key : baselayout ] [ Applications found : 1 ] * sys-apps/baselayout Latest version available: 1.8.3 Latest version installed: 1.8.3 Homepage: http://www.gentoo.org/ Description: Base layout for Gentoo Linux filesystem (incl. initscripts and sysvinit) I think this is the reason why we still get duplicates of this bug.
azarah@nosferatu profiles $ cvs update U package.mask azarah@nosferatu profiles $ grep baselayout package.mask # it in baselayout, but we need >=gawk-3.1.0-r3 #we no longer use sysvinit (rolled into baselayout) azarah@nosferatu profiles $ emerge -s baselayout Searching... [ Results for search key : baselayout ] [ Applications found : 1 ] * sys-apps/baselayout Latest version available: 1.8.4 Latest version installed: 1.8.4 Homepage: http://www.gentoo.org/ Description: Base layout for Gentoo Linux filesystem (incl. initscripts and sysvinit) azarah@nosferatu profiles $ I suggest you change your rsync server, or cleanout /usr/portage, and do a emerge rsync again. This bug is fixed. If you have rsync problems, please open another bug for that.
For myself this bug is fixed since my first comment, but i'm wondering why i still get mails for this bug... and after some searching hear is the reason: the baselayout-1.8.4.ebuild is marked for unstable KEYWORDS="~x86 ~ppc ~sparc ~sparc64 ~alpha" so maybe moving this ebuild to stable will finaly close this bug.
*** Bug 9913 has been marked as a duplicate of this bug. ***
I still have the error... --- BEGIN SHELL --- jupiter init.d # emerge -s baselayout Searching... [ Results for search key : baselayout ] [ Applications found : 1 ] * sys-apps/baselayout Latest version available: 1.8.4.1 Latest version installed: 1.8.4.1 Homepage: http://www.gentoo.org/ Description: Base layout for Gentoo Linux filesystem (incl. initscripts and sysvinit) jupiter init.d # rep need /etc/init.d/netmount local myneed="net" local myneed="${myneed} portmap" need ${mydeps} jupiter init.d # grep Header netmount # $Header: /home/cvsroot/gentoo-src/rc-scripts/init.d/netmount,v 1.17 2002/08/25 20:29:07 azarah Exp $ jupiter init.d # --- END SHELL ---
nosferatu init.d # grep Header netmount # $Header: /home/cvsroot/gentoo-src/rc-scripts/init.d/netmount,v 1.19 2002/10/13 19:20:23 azarah Exp $ nosferatu init.d # -- You do not know how to update config files. # emerge --help config
I do know how to update config files. jupiter init.d # find /etc -name ._'*' jupiter init.d # Okay ... I believe it is fixed, but not on my system. I emerged baselayout again, but the netmount was still version 1.17. It did say the following: sys-apps/baselayout selected: none protected: 1.7.8-r1 1.8.4.1 omitted: none So I thought lets delete the /etc/init.d/netmount and re-emerge baselayout. That worked now the netmount is version 1.19. But this can't be the thing to do, or is it ? Who is holding on to version 1.7.8-r1 ?
*** Bug 10375 has been marked as a duplicate of this bug. ***
*** Bug 10488 has been marked as a duplicate of this bug. ***
*** Bug 10682 has been marked as a duplicate of this bug. ***
this has started happening again for me - not quite sure why.. first, I've checked the awk part of the script detects my nfs mount correctly which should theoretically enable portmap as a dependancy root # awk '!/^#/ && $3 == "nfs" && $4 !~ /noauto/ { print $0 }' /etc/fstab tremor:/usr/portage/distfiles /usr/portage/distfiles nfs noatime 0 0 so that definately finds something.. next, # Only try to mount NFS filesystems if portmap was started. well, portmap didn't start therefore the nfs mount doesn't happen - I'm guessing this is a problem in the depend section but it's not staring me in the face at 1am :) tell me if I've done something dumb, (and yes, portmap is definately added in runlevel default not that it should matter if I'm reading this correctly). depend() { local myneed="net" local myuse="" # # Only have Portmap as a dependency if there is a nfs mount in fstab # that should be mounted at boot time. Also filter out comments. # local nfsmounts="$(awk '!/^#/ && $3 == "nfs" && $4 !~ /noauto/ { print $0 }' /etc/fstab)" if [ -n "${nfsmounts}" ] then local myneed="${myneed} portmap" local myuse="${myuse} nfs" fi need ${myneed} use ${myuse} }
Err, I know why. Its prob because im an idiot :/ Check netmount's cache from /mnt/.init.d/depcache: ---- #*** /etc/init.d/netmount *** myservice="/etc/init.d/netmount" myservice="${myservice##*/}" echo "RCSCRIPT ${myservice}" depend() { return 0 } depend() { local myneed="net" local myuse="" if [ -n "${nfsmounts}" ] then local myneed="${myneed} portmap" local myuse="${myuse} nfs" fi need ${myneed} use ${myuse} } depend ----- It ... err ... filter out the awk command :/
Ok, its a one char fix .. change in /lib/rcscripts/awk/cachedepends.awk: ---- # Filter out comments and only process if its a rcscript if (($0 !~ /[[:space:]]*#/) && (ISRCSCRIPT == 1 )) { ---- To: --- # Filter out comments and only process if its a rcscript if (($0 !~ /^[[:space:]]*#/) && (ISRCSCRIPT == 1 )) { --- Note the '^' in front of '[[:space:]]' ....
yep, that did it, thanks! (also removed duplicate cc (I was the original reporter, just with a different address at the time.))