Sun SPARCstorage RSM Array 2000
 
Sun Disk, 500 GB, 7200 rpm, SATA
 
Sun Storage 6540 Array
 
Sun Storage 3510 FC Array
 
Sun Storage MultiPack FC Desktop Array
 
Sun Storage Flexline 280 Array
 
Sun Storage T3 Array
 
Sun Storage T3+ Array
 
Exadata Database Machine X2-2 Hardware
 
Sun Netra st D1000 Array
 
Sun Storage A3000 Array
 
Sun Storage 2510 Array
 
Sun Storage UniPack Disk Drive
 
Sun Storage J4200 Array
 
Sun Flash F5100 Array
 
Sun Storage J4650 Array
 
Sun Storage A5200 Array
 
Sun Storage D1000 Array
 
Sun Storage D2 Array
 
Sun Storage 9970 System
 
Sun Blade 6000 Disk Module
 
Sun Storage 2530 Array
 
Sun Storage 2540 Array
 
Sun Storage 3310 Array
 
Sun Storage A5000 Array
 
Sun Storage MultiPack SCSI Desktop Array
 
Sun Storage D240 (StorEdge) Media Tray
 
Sun Storage 6130 Array
 
Sun Storage 6140 Array
 
Sun Storage A3500 SCSI Array
 
Sun Storage Flexline 240 Array
 
Sun Netra st A1000 Array
 
Sun Netra st D130 Storage Array (DC)
 
Sun Storage 6920 System
 
Sun Storage 9990 System
 
Sun Storage Flexline 380 Array
 
Sun Storage 3511 SATA Array
 
Sun Storage 9960 System
 
Sun Storage 9985 System
 
Sun Storage J4500 Array
 
Sun Storage 9980 System
 
Sun Storage A3500 FC Array
 
Oracle Solaris Express
 
Sun Storage 3120 SCSI Array
 
Sun Storage 3320 SCSI Array
 

GCS>Sun Microsystems>Storage - Disk>Drives - Misc
 

Applies to:

Goal

# format
Searching for disks...done

AVAILABLE DISK SELECTIONS:
0. c0t0d0 <drive not available: formatting>
/sbus@6,0/SUNW,socal@d,10000/sf@0,0/ssd@w210000163787753a,0

Solution

With older disk drives which did not perform automatic bad block reallocation, the Solaris format(1m) utility was used to mark bad blocks and to format disk drives to avoid using those bad blocks.

With modern disk drives, the format(1m) utility is only very, very rarely used to actually perform a format of a disk drive, mainly because modern disk drives perform auto-reallocation of bad blocks (and therefore automatically "retire" that defective piece of media) upon certain read or write errors, without the disk drive needing to be completely re-formatted.

Due to the length of time which a disk drive requires to perform a "Format Unit" command (e.g. some hours, depending on the disk drive capacity and performance), the format(1m) utility is designed to show when a disk drive is formatting, and to allow the user to perform actions against other disk drives during that time.

This is why the <drive not available: formatting> message was introduced, to indicate which disks were not available for the format(1m) utility to use, at that particular time.

During formatting (e.g. while it is executing a "Format Unit" command), a disk drive will respond to an "Inquiry" command, but will return an error for almost every other command that is sent to it, including a "Read" command. Typically it will return a "check condition" to the host, with sense key 0x2 (not ready), ASC 0x4, ASCQ 0x4 (logical unit not ready, format in progress).

The format(1m) utility will therefore show the message:

<drive not available: formatting>

when the disk drive has responded to an "Inquiry" command, but has returned an error when the utility then attempts to read the label from LBA 0 on the disk. Note: The contents of the disk label are not interpreted at this stage - format(1m) just checks that the "Read" command to LBA 0 is successful. This successful "Inquiry" command, followed by an unsuccessful "Read" command, is the behavior of a disk drive which is in the process of formatting.

However, if the user has not started to format a disk drive, and sees this <drive not available: formatting> message, then it is possible that the disk drive is faulty.

The typical cause is that the disk drive is reporting a "not ready" sense key due to an internal fault. Disk drives in that faulty condition often respond to an "Inquiry", but are unable to respond successfully to a "Read" command - which is very similar behavior to a disk drive while it is performing a "Format Unit" command.

Other less likely causes could include:

• The disk drive is reporting an unrecoverable read error on LBA 0; or
• The disk drive has just been power-cycled or reset, it is in the process of becoming ready, and will become ready (and able to respond to a "Read" command) within a few seconds.

The format(1m) utility sends all commands to disk drives with the USCSI_SILENT flag set, so no error messages are logged in /var/adm/messages with the default value of sd_error_level (ssd_error_level when using the ssd driver) even if errors do occur. This is to prevent customer concern when some expected errors occur, but this makes it difficult to find the exact cause of the <drive not available: formatting> message.

In some cases, running:

# format -Mm

may be helpful as this reports additional messages about what is happening.

That message may sometimes be reported by format(1m) for a LUN from a disk array, for similar reasons - format(1m) treats a LUN from a disk array just like an individual disk drive. So if a customer runs the format(1m) utility and that disk array LUN responds correctly to an "Inquiry" command, but reports an error for a "Read" command, the same message will be seen from format(1m).

This may occur due to a temporary condition e.g. when the array has just been powered-on or reset. In that case, disk arrays tend to respond quickly to "Inquiry" commands from a host, but they report a sense key of "not ready" for "Read" commands until all of their internal startup processing has been done (which can take a few minutes). If this message from format(1m) is seen from a disk array LUN, even when the disk array has not recently been powered-on or reset, then this requires further investigation on the disk array for possible problems.

This solution has no attachment