Comment 5 for bug 1463046

Revision history for this message
Mathieu Trudel-Lapierre (cyphermox) wrote :

The devices don't appear to have the same serial or even the same product name, I'm unsure why they're be considered as multipathed even with the apparent same scsi_id as you've discovered.

What does multipath -ll output on this system? Also, could you please run the following commands:

for d in /dev/sda /dev/sdb; do echo "page 0x80: $d:'$(sudo /lib/udev/scsi_id --whitelisted --page=0x80 --device=$d)'"; done
for d in /dev/sda /dev/sdb; do echo "page 0x83: $d:'$(sudo /lib/udev/scsi_id --whitelisted --page=0x83 --device=$d)'"; done
for d in /dev/sda /dev/sdb; do echo "page pre: $d:'$(sudo /lib/udev/scsi_id --whitelisted --page=pre-spc3-83 --device=$d)'"; done

In case we're dealing here with non-standard devices, I'd like to whether there is an alternative method we can use that would get better results. We've seen here that page 0x83 was used, but we should find out whether 0x80 is available, or if this also needs kernel/driver work.