On May 19, 2011, at 12:55 AM, Evaldas Auryla wrote:

> Hi, we have SunFire X4140 connected to Dell MD1220 SAS enclosure, single 
> path, MPxIO disabled, via LSI SAS9200-8e HBA. Disks are visible with 
> sas-addresses such as this in "zpool status" output:
> 
>        NAME                       STATE     READ WRITE CKSUM
>        cuve                       ONLINE       0     0     0
>          mirror-0                 ONLINE       0     0     0
>            c9t5000C50025D5AF66d0  ONLINE       0     0     0

This really isn't the SAS address, but it is similar.

>            c9t5000C50025E5A85Ad0  ONLINE       0     0     0
>          mirror-1                 ONLINE       0     0     0
>            c9t5000C50025D591BEd0  ONLINE       0     0     0
>            c9t5000C50025E1BD56d0  ONLINE       0     0     0
>   ...
> 
> Is there an easy way to map these sas-addresses to the physical disks in 
> enclosure ?

No. But there are several ways to map them. The easiest is to get the disk
serial number and cross-reference to the fmtopo output. I've got a script that
does this for NexentaStor. Perhaps someone can port to Solaris (NexentaStor
has a utility, hdddiso, that provides disk info, such as the serial number, 
that is 
easily parseable in a script :-)

fmtopo can talk to many enclosures (run with root privileges) and can get the 
mappings.  A typical line looks like:
hc://:product-id=QUANTA-Storage-JB7:server-id=:chassis-id=50016360001d0008:serial=WD-WMAY00547508:part=ATA-WDC-WD2003FYYS-0:revision=1D01/ses-enclosure=2/bay=21/disk=0

If you have more than one enclosure, each enclosure is enumerated, but the 
cross-reference needs to be done elsewhere (hint: use chassis-id)

In these cases, there is an assumption that the "bay=21" logical representation
correlates to the silk-screened labels on the enclosure. This assumption can 
trip
you up.
 -- richard

_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Reply via email to