On Thursday February 14, ian@stams.strath.ac.uk wrote: > > ## Q.1: Why does the first device get listed as "sdb" and not "sdb1" ? > Note that "mdctl -E /dev/sdb1" doesn't work, it requires "/dev/sdb" > unlike the c and d drives...did I mess up? Yep, you messed up. You will see the same discrepancy in /proc/mdstat. Not that there is much harm in this. Most of my RAID array are build from whole disks, not partitions. It means that autodetect cannot work (because there is no partition ID) but regular readers will know what I think of auto-detect. > > >From /proc/scsi/scsi: > > Host: scsi2 Channel: 00 Id: 00 Lun: 00 > Vendor: SEAGATE Model: ST336704LC Rev: 0004 > Type: Direct-Access ANSI SCSI revision: 03 > Host: scsi2 Channel: 00 Id: 01 Lun: 00 > Vendor: SEAGATE Model: ST336704LC Rev: 0004 > Type: Direct-Access ANSI SCSI revision: 03 > Host: scsi2 Channel: 00 Id: 02 Lun: 00 > Vendor: IBM Model: DDYS-T36950M Rev: S93E > Type: Direct-Access ANSI SCSI revision: 03 > > > ## Q.2: The failed device /dev/sdd1 corresponds to the IBM disk, > ## yes? Maybe.. sdd should correspond the the fourth drives listed in /proc/scsi/scsi, but you have only listed 3. Maybe you didn't include a full listing. > (If so, this is the third Deskstar drive to fail in the array, the > 2 SEAGATEs are replacement drives!) Deskstars seem to be like that. We have had several fail. I doubt I'll be buying drives from IBM for a while. > ## Q.3: Can someone verify that to replace the drive I do: > > mdctl --remove /dev/md0 /dev/sdd1 > < hot swap failed drive with new one> You do know that scsi hot-swap if not fully supported by Linux?!? You will need to do something like: echo 'scsi remove-single-device 2 0 2 0' > /proc/scsi/scsi # pull out the old # push in the new echo 'scsi add-single-device 2 0 2 0' > /proc/scsi/scsi It will probably work, but no guarantees > mdctl --add /dev/md0 /dev/sdd1 > mdctl --assemble /dev/md0 /dev/sdb /dev/sdc1 /dev/sdd1 > (note - use sdb, not sdb1?) You do not need --assemble, it is already assembled. After the mdctl --add /dev/md0 /dev/sdd1 it will start rebuilding (you can watch in /proc/mdstat). You would only need --assemble if you shut the array down (--stop) and wanted to restart it, such as when you reboot. NeilBrown - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
