Hello > > # cat /proc/mounts > ::: > > aufs / aufs = > > > rw,si=3Dcf986180,xino=3D/mnt/rwdisk/.aufs.xino,nowarn_perm,br: > /mnt/rwdisk= > > =3Drw:/mnt/overlay=3Dro:/mnt/rootfs=3Dro 0 0 > > So you didn't specify udba=inotify option? > I thought that you did because CONFIG_AUFS_HINOTIFY was > enabled in your > configuration.
No, unfortunately not. I didn't do that, because of the possible performance problems. But see below... > It is strongly recommended to specify udba=inotify option when you > modify branches directly (bypassing aufs). > In your case, > host # mount -o remount,udba=inotify / > installer # rsync (and update the branch on 'host') > host # mount -o remount,udba=reval / Ah, it's ok to set inotify only during the direct branch access? I will try that. Thank you for this hint. > If everything goes well, you may not need to copy files manually after > rsync since new files will be visible. But when the file > already existed > on the upper branch and you want to see the one on the lower > branch, you > may still need to copy it from the lower branch. In that case I have the problem, that a further direct branch access won't update the file again. What is the safe way to synchronize the union with the branch? I have tried the following, but I'm not sure if it's dangerous: # rm /<path-to-file> # remove the file in the union # rm /mnt/rwdisk/<path-to-.wh-file> # remove the .wh-file in the read-writable partition After that, the lower branch file is in the union again. Is there a better method? (But this was not the procedure which caused the segmentation fault!) > > You mean "chattr +i <file>"? > > I didn't set the flag on any file. Is it possible that a daemon is = > > setting some flags? > > Although I don't know such deamon exist or not, I want you to try > $ lsattr /mnt/rwdisk/the_file_you_were_going_to_copy > > If it shows the file is immutable, I can understand what the > bug is and > fix the problem. Unfortunatly I don't know which file causes the segmentation fault. In my rsync-Script I didn't have verbose output. :-( > Otherwise, I have to spend many nights unsleep. :-) Oh, no, no, enough sleep is more important! I will change my scripts to set "inotify" during updates. Thanks for your help. Best regards Elmar
smime.p7s
Description: S/MIME cryptographic signature
------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/
