Talk:SRE/Dc-operations/Platform-specific documentation/Sun Fire X4500 and X4540

Rendered with Parsoid
From Wikitech

For the X4500

We were trying to work out which disks were which recently when we had a couple of failures (under Linux). The mad disks were reported as /dev/sdab and /dev/sdae. Hopping on the box, I saw:

# lspci | grep Marvell
 01:01.0 SCSI storage controller: Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller (rev 09)
 02:01.0 SCSI storage controller: Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller (rev 09)
 05:01.0 SCSI storage controller: Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller (rev 09)
 06:01.0 SCSI storage controller: Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller (rev 09)
 0a:01.0 SCSI storage controller: Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller (rev 09)
 0b:01.0 SCSI storage controller: Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller (rev 09)

That's how they show up. If the order is the same as in Solaris, then...

# ls -l /dev/disk/by-path/pci-0000\:06*
 lrwxrwxrwx 1 root root  9 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-0:0:0:0 -> ../../sdy
 lrwxrwxrwx 1 root root 10 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-0:0:0:0-part1 -> ../../sdy1
 lrwxrwxrwx 1 root root  9 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-1:0:0:0 -> ../../sdz
 lrwxrwxrwx 1 root root 10 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-2:0:0:0 -> ../../sdaa
 lrwxrwxrwx 1 root root 10 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-3:0:0:0 -> ../../sdab
 lrwxrwxrwx 1 root root 10 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-4:0:0:0 -> ../../sdac
 lrwxrwxrwx 1 root root 11 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-4:0:0:0-part1 -> ../../sdac1
 lrwxrwxrwx 1 root root 10 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-5:0:0:0 -> ../../sdad
 lrwxrwxrwx 1 root root 10 2009-04-15 12:19 /dev/disk/by-path/pci-0000:06:01.0-scsi-7:0:0:0 -> ../../sdaf

Our bad disk /dev/sdab would be the 4th disk off the 4th controller. And, it seemd like pci-0000:06:01.0-scsi-6:0:0:0 "ought to be" sdae which would make that the 7th disk off that controller.

The boot disks would have been the 4th controller as well (disk 1 and 5 of of same controller have md0 ( = sdy1 and sdac1), and md0 has slash). That jives with the diagram.

This means, maybe: the bad disks were disks 36 and 25 on the diagram.