Top posting to keep flow... Sorry Benjamin, I think I missed something really obvious here. I didn't read your original message carefully enough, and thought you were having a problem with the kernel autodetection of your RAID. The part I missed is the "(skipped)" message you get...
Looking through the /etc/init.d/checkfs script, it appears that the "(skipped)" message will occur if you have "noauto" specified for the filesystem in /etc/fstab. Take that out, and I will bet everything will work just like it used to. If you _do_ decide to rebuild the raid array, and you want the kernel to autodetect things when it starts, then you need to have the right partition types and create the array with the "persistent-superblock" option, or use mdadm to create the array. This is nice, because then you don't need a raidtab at all... -Richard Benjamin Sobotta wrote: >Hi! > >I think I'll redo the RAID as soon as I find time. For now it'll do the way it >is. As long as it works... I really can't mess with it right now since I >really need the machine. Maybe this weekend... > >Thanks guys! > >Ben > >On Wednesday 25 May 2005 19:50, Richard Fish wrote: > > >>Christoph Gysin wrote: >> >> >>>Benjamin Sobotta wrote: >>> >>> >>>>I set up a software RAID1 with two SATA disks. /home resides on the raid. >>>>This used to work great for month. At boot the machine would test the >>>>raid and start it. Great! >>>>2 days ago however I put loop-AES on top of it in order to encrypt my >>>>home directory. Since then it still finds the raid at startup, however, >>>>doesn't start it anymore. (displays "skipping"). I always have to start >>>>by hand via "raidstart /dev/md0". Then it works fine. >>>>Can someone tell me why it stopped to start automatically - or even >>>>better how to get it back to work?! >>>> >>>> >>>The raid configuration (as defined in the raidtab) gets stored in a >>>superblock on each disk. If the partitiontype is fd, linux reads the >>>superblock at boot and automatically starts the array. >>> >>> >>All true.... >> >> >> >>>Since you have now encrypted the whole device, linux can't read the >>>superblock at boot. raidstart works, because it reads /etc/raidtab and >>>sets up the array. >>> >>> >>Sorry, this is all false. Benjamin has placed the encypted volume *on >>top* of the raid array, not underneath it. Besides if you clobber the >>superblocks, and your raidtab matches your actual configuration (has >>"persistent-superblock 1"), raidstart will throw an error. >> >>The array is not autodetected because the kernel cannot decipher his >>partition table at boot time. >> >>-Richard >> >> -- [email protected] mailing list
