Once you have created a new volume in the CLI tool, Linux does not automatically see the disks. To remedy, use the following:
# echo “- – -” > /sys/class/scsi_host/host0/scan
If one of the volumes was still initializing when you did that, it will have been detected by the driver, but not assigned a device node. In that case, performing the above ritual after it has been initialized will not help. You first have to remove the device again, then rescan again:
# dmesg|grep 0:0:
scsi 0:0:0:3: Direct-Access Areca ARC-1260-VOL#03 R001 PQ: 1 ANSI: 3
scsi 0:0:0:3: Attached scsi generic sg4 type 0
# echo “scsi remove-single-device 0:0:0:3” > /proc/scsi/scsi
# echo “- – -” > /sys/class/scsi_host/host0/scan
# dmesg|grep 0:0:
scsi 0:0:0:3: Direct-Access Areca ARC-1260-VOL#03 R001 PQ: 1 ANSI: 3
scsi 0:0:0:3: Attached scsi generic sg4 type 0
scsi 0:0:0:3: Direct-Access Areca ARC-1260-VOL#03 R001 PQ: 0 ANSI: 3
sd 0:0:0:3: Attached scsi disk sdd
sd 0:0:0:3: Attached scsi generic sg4 type 0
Incoming Links (via Technorati):
Nothing Reported
July 8th, 2009 at 15:50
I wonder whether “udevadm trigger” (formerly udevtrigger) and/or maybe partprobe would do the trick?
July 8th, 2009 at 15:57
Hmm, I don’t know either, but do they work at such a low level? Partprobe will find new partitions on devices, but I don’t think it would find new devices themselves…
July 8th, 2009 at 16:11
You’re probably right about partprobe.
The man page of udevadm however states that the trigger action “requests device events from the kernel”.
I’ll need to check that next time I setup an Areca based system 🙂