Matt Connolly
2010-Jun-30 21:16 UTC
[zfs-discuss] Zpool mirror fail testing - odd resilver behaviour after reconnect
I have a Opensolaris snv_134 machine with 2 x 1.5TB drives. One is a Samsung Silencer the other is a dreaded Western Digital Green. I''m testing the mirror for failure by simply yanking out the SATA cable while the machine is running. The system never skips a beat, which is great. But the reconnect behaviour is vastly different on the two drives. 1. Samsung reconnect. `cfgadm` reported the drive as connected but unconfigured. After running `cfgadm -c configure sata1/1`, the drive automatically came online in the zpool mirror and resilvered its differences which completed in about 10 seconds. This excellent. 2. WD Green reconnect. `cfgadm` reported the drive as disconnected. I had to use the ''-f'' option to connect the drive and then configure it: matt at vault:~$ cfgadm Ap_Id Type Receptacle Occupant Condition sata1/0 sata-port disconnected unconfigured failed sata1/1::dsk/c8t1d0 disk connected configured ok matt at vault:~$ pfexec cfgadm -c connect sata1/0 cfgadm: Insufficient condition matt at vault:~$ pfexec cfgadm -f -c connect sata1/0 Activate the port: /devices/pci at 0,0/pci8086,4f4d at 1f,2:0 This operation will enable activity on the SATA port Continue (yes/no)? yes matt at vault:~$ cfgadm Ap_Id Type Receptacle Occupant Condition sata1/0 disk connected unconfigured unknown sata1/1::dsk/c8t1d0 disk connected configured ok matt at vault:~$ pfexec cfgadm -c configure sata1/0 matt at vault:~$ cfgadm Ap_Id Type Receptacle Occupant Condition sata1/0::dsk/c8t0d0 disk connected configured ok sata1/1::dsk/c8t1d0 disk connected configured ok After this point, zpool resilvered the entire 243GB dataset. I suspect that the automatic connect is simply a firmware problem and yet another reason to NOT BUY Western Digital Green drives. But my real question is: Why does zpool want to resilver the entire dataset on drive, but not the other?? -- This message posted from opensolaris.org