Date: Thu, 30 May 2013 15:43:13 -0400 From: Douglas Gilbert <dgilbert@interlog.com> To: Riccardo Veraldi <riccardo.veraldi@gmail.com> Cc: freebsd-scsi@freebsd.org Subject: Re: troubles with Seagate (EMC2) disk Message-ID: <51A7ABD1.7090706@interlog.com> In-Reply-To: <51A77152.8030100@gmail.com> References: <51A72605.4050303@gmail.com> <51A76794.2080500@interlog.com> <51A77152.8030100@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 13-05-30 11:33 AM, Riccardo Veraldi wrote: > thank you very much! > > your utility is outstanding!! > > Format in progress, 5.100% done > Format in progress, 6.99% done > Format in progress, 6.99% done > Format in progress, 6.99% done > Format in progress, 7.99% done > Format in progress, 7.99% done > Format in progress, 8.100% done > Format in progress, 8.100% done > Format in progress, 8.100% done That looks a bit noisy, I don't see that in Linux. Did the progress indication really go backwards? > anyway for mistake I did not issued a --wait command... Best not to use wait, as your OS needs to wait an awfully long time for that FORMAT UNIT command to complete. Some OSes get fidgety, and send some sort of reset to the disk which they think is in some degenerate state. I have a 3 TB Seagate disk that took over 13 hours to format (in my case I was adding protection information which expands the actual sector size from 512 to 520 bytes). > I am formatting a 512GB seagate disk 512 bytes per sector. > > how long is the default timeout ? sg_format currently chooses 20 seconds if --wait is not given (which means IMMED=1 in FORMAT UNIT) and 15 hours otherwise! With 4 TB disks, that probably should be 20 hours. Silly. Ask Seagate/EMC about how long your disk will take to format. As I said best not to use --wait; then the FORMAT UNIT is sent with the IMMED bit set and sg_format polls the disk checking for the progress indication with TEST UNIT READY (or REQUEST SENSE). > if timeout occours I think I have to restart formatting again :( Oh yes. If FreeBSD is anything like Linux, then timeouts on SCSI commands are best avoided. Especially bad on FORMAT UNIT, then you can read about "format corrupt" state while you wait for the format to run again. Doug Gilbert > On 5/30/13 4:52 PM, Douglas Gilbert wrote: >> On 13-05-30 06:12 AM, Riccardo Veraldi wrote: >>> Hello, >>> I have a Seagate ST350000 CLAR500 disk. >>> >>> it is attached to my Sun Blade 1000 running FreeBSD 9.1 >>> >>> I need to label it, create partition and move te root /usr/ var and swap from >>> the main boot disk which has corrupted blocks. >>> >>> so this disk will become my boot disk. >>> >>> the problem is that I cannot partition it or even access it in anyway >>> >>> dd: /dev/da6: Invalid argument >>> 1+0 records in >>> 0+0 records out >>> 0 bytes transferred in 0.000700 secs (0 bytes/sec) >>> >>> this is the disk information: >>> >>> da6 at isp0 bus 0 scbus2 target 1 lun 0 >>> da6: <SEAGATE ST350000 CLAR500 FDBD> Fixed Direct Access SCSI-3 device >>> da6: 100.000MB/s transfers WWNN 0x2000001862133c8d WWPN 0x2100001862133c8d >>> PortID 0xe4 >>> da6: Command Queueing enabled >>> da6: 476940MB (961745856 520 byte sectors: 255H 63S/T 59865C) >>> >>> the problem is that disk byte sectors are 520 and not 512 >>> >>> is there a way to reformat it to 512 byte sectors ? >>> >>> is there a way to make it usable ? >>> >>> the SCSI III disk has a FC - AL interface (Internal Sun Blade 1000 Caddy). >> >> In my sg3_utils package (see http://sg.danny.cz/sg/sg3_utils.html ) >> there is a utility called sg_format. If you are sure >> that you do not want the existing data on that disk, try: >> sg_format --format --size=512 /dev/da6 >> >> >> Doug Gilbert > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51A7ABD1.7090706>