Thans Wol and Nikhil. On Sat, Apr 28, 2018 at 12:52 AM, Wols Lists <antlists@xxxxxxxxxxxxxxx> wrote: > On 28/04/18 05:32, Nick Leli wrote: >> These (sd[ab]) drives are not explicitly mounted (at least in >> /proc/mounts), yet this seems to cause a problem. So I'm curious: >> >> 1. Is it possible to mount these devices, while a RAID array is active? >> 2. If mounting is not possible, is there another way to verify that >> each disk contains the correct data (besides /proc/mdstats or dd)? > > There's nothing there to mount! > > The only time you can mount a raid member, is if you've got a > file-system directly on a v0.9 or v1.0 mirror. You've got a v1.2. To add > to that, as I understand you, you have an LVM volume on your raid array. > > So sd5 contains a raid "whatever-it-is", which then has a lvm > "whatever-it-is" starting at some random position within the raid > whatever-it-is, which then itself has your file-system at some random > position within that! > > Seeing as mount only works with file-systems, good luck in working out > where the file-system is physically located to mount it! Not knowing > LVM, it wouldn't surprise me to find it's not actually a contiguous > chunk of disk, so mounting it would be extremely dangerous. > > Oh - and that md5 sum idea *might* work, but the more layers you have in > the stack the more likely it is to fail, and any management work like > replacing disks is almost certain to mess it up. > > I'm afraid you'll really have to assume that if all the monitoring tools > say every layer is working correctly as designed, that the whole system > is okay. > > Cheers, > Wol -- 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
