Bug#543375: Info received (Bug not solved)
2009/10/21 Gabriele Giacone losga...@libero.it Ok, solved. My boot/shutdown sequences were wrong. I reinstalled every package that owns a script under /etc/init.d. Not a simple reinstall. Before each reinstall, I did a update-rc.d -f script remove because, correct me if I'm wrong, if rcN links already exist, installation won't correct them, unless specified in preinst/postinst scripts. Now, in util-linux.postinst, for hwclock.sh and hwclockfirst.sh, there's a -f remove before start and it's ok. Reinstalling initscripts and many other packages, doing this cleanup, sequence numbers have significantly changed. Under /etc/rc6.d, before after S01sendsigsS20sendsigs S01umountnfs.shS31umountnfs.sh S01umountfsS40umountfs S01umountrootS60umountroot S01rebootS90reboot After fixing the scripts above, system booted fine. Then I fixed all the remaining. Hello, would it be enough to just manually correct symlinks as described above without reinstalling anything? I wouldn't like to render my system unusable because of messing with those critical packages. Besides, aren't any changes required in other rcX.d - is rc6.d enough? And what did you exactly mean by Then I fixed all the remaining? If you have corrected the problem could you please describe step by step what should be done to fix this? B. Regards, Tomasz Grzelak
Bug#543375: Info received (Bug not solved)
Tomasz Grzelak wrote: 2009/10/21 Gabriele Giacone losga...@libero.it mailto:losga...@libero.it [...] Under /etc/rc6.d, before after S01sendsigs S20sendsigs S01umountnfs.sh S31umountnfs.sh S01umountfs S40umountfs S01umountroot S60umountroot S01reboot S90reboot After fixing the scripts above, system booted fine. Then I fixed all the remaining. Hello, Hi Tomasz would it be enough to just manually correct symlinks as described above without reinstalling anything? I wouldn't like to render my system unusable because of messing with those critical packages. Besides, aren't any changes required in other rcX.d - is rc6.d enough? And what did you exactly mean by Then I fixed all the remaining? If you have corrected the problem could you please describe step by step what should be done to fix this? I started to modify the above ones then I continued with all the other ones. My goal was to get the official sequence number. Steps? For each init.d script, remove current rcN links and reinstall the package it belongs. Use the following script. It let you choose among /etc/init.d scripts and the reinstallation of the package they belong. Package reinstall implies reconfigure? My system is a desktop, I didn't do complex reconfiguration. Compare before and after sequence numbers and you should see a lot of differences. --- for i in /etc/init.d/* do _file=$(basename $i) echo -e File $_file OK? \c read _answ [ $_answ != y ] continue ### find to see current sequence numbers find /etc/rc*|grep $_file ### removing current links update-rc.d -f $_file remove ### from file to package dpkg -S $i _pkg=$(dpkg -S $i|cut -d: -f1) ### reinstall it? echo -e Reinstall package $_pkg OK? \c read _answ [ $_answ != y ] continue apt-get -y --reinstall install $_pkg ### find to see new sequence numbers find /etc/rc*|grep $_file done --- All of these fixes should be done by package upgrade phase but in the most cases it doesn't happen. I've installed this squeeze a couple of months ago and during this period all of these sequence number have changed? G -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#543375: Info received (Bug not solved)
2009/10/21 Gabriele Giacone losga...@libero.it I started to modify the above ones then I continued with all the other ones. My goal was to get the official sequence number. Steps? For each init.d script, remove current rcN links and reinstall the package it belongs. Use the following script. It let you choose among /etc/init.d scripts and the reinstallation of the package they belong. Package reinstall implies reconfigure? My system is a desktop, I didn't do complex reconfiguration. Compare before and after sequence numbers and you should see a lot of differences. --- for i in /etc/init.d/* do _file=$(basename $i) echo -e File $_file OK? \c read _answ [ $_answ != y ] continue ### find to see current sequence numbers find /etc/rc*|grep $_file ### removing current links update-rc.d -f $_file remove ### from file to package dpkg -S $i _pkg=$(dpkg -S $i|cut -d: -f1) ### reinstall it? echo -e Reinstall package $_pkg OK? \c read _answ [ $_answ != y ] continue apt-get -y --reinstall install $_pkg ### find to see new sequence numbers find /etc/rc*|grep $_file done --- Hello! Thank you very much for your description the script. Actions to take look much clearer now for me. I'll try to 'implement' them in next few days. I'll let you know how it goes. B. Regards, Tomasz Grzelak
Bug#543375: Info received (Bug not solved)
Ok, solved. My boot/shutdown sequences were wrong. I reinstalled every package that owns a script under /etc/init.d. Not a simple reinstall. Before each reinstall, I did a update-rc.d -f script remove because, correct me if I'm wrong, if rcN links already exist, installation won't correct them, unless specified in preinst/postinst scripts. Now, in util-linux.postinst, for hwclock.sh and hwclockfirst.sh, there's a -f remove before start and it's ok. Reinstalling initscripts and many other packages, doing this cleanup, sequence numbers have significantly changed. Under /etc/rc6.d, before after S01sendsigsS20sendsigs S01umountnfs.shS31umountnfs.sh S01umountfsS40umountfs S01umountrootS60umountroot S01rebootS90reboot After fixing the scripts above, system booted fine. Then I fixed all the remaining. When is the last time the superblock is written? Isn't it when filesystem is umounted/mounted read-only on shutdown? Could it be UTC time when it's umounted? G. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
