You'd think something could be put in the RAID superblocks indicating precendence. Or smarts in general of 'if a component array of the array is not assembled, assemble it first'. Then those that just want to boot to those kinds of RAID arrays without managing init.rd stuff can do so. Derek On 4/17/05, TJ <systemloc@xxxxxxxxxxxxx> wrote: > I searched through the mailing list archive but I can not find this. I > remember a discussion about trouble being caused from lack of control of the > order that kernel autodetect chooses, and a patch to be able to control it. > I'm running an array on top of another array right now, a RAID 5 on a linear. > I understand the consequences of this, it's temporary. Still, the kernel > tries to bring up the RAID 5 array before the linear, so the linear array is > not included. If there is no way to correct this in the kernel, I need some > help on understanding how init.rd level raid starting is supposed to work. It > must happen very early, obviously, to ensure that auto-fsck works correctly. > I'm guessing that it should be first, or close to first in the scripts. The > general idea would be simply to run mdadm from a script to start the arrays, > then? > > Thanks > TJ > - > To unsubscribe from this list: send the line "unsubscribe linux-raid" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html > -- Derek Piper - derek.piper@xxxxxxxxx http://doofer.org/ - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
