Date: Tue, 11 Mar 2014 17:09:23 +0100 From: Willem Jan Withagen <wjw@digiware.nl> To: "Pokala, Ravi" <rpokala@panasas.com>, "hardware@freebsd.org" <hardware@freebsd.org> Subject: Re: Differences between disk controllers Message-ID: <531F3533.7000407@digiware.nl> In-Reply-To: <CF4474B3.108A9A%rpokala@panasas.com> References: <CF4474B3.108A9A%rpokala@panasas.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 11-3-2014 16:34, Pokala, Ravi wrote: > > >> Date: Mon, 10 Mar 2014 22:52:16 +0100 >> From: Willem Jan Withagen <wjw@digiware.nl> >> To: hardware@freebsd.org >> Subject: Differences between disk controllers >> Message-ID: <531E3410.5090602@digiware.nl> >> Content-Type: text/plain; charset=ISO-8859-1 >> >> Hi, >> >> Running: >> FreeBSD 10.0-STABLE #0 r262730 >> >> I've build a ZFS system, with 4 disks on a regular Marvell controller, >> and 4 disks on an Areca controller. Disk are 4T WD REDs. >> >> I noticed that there were some discrepancies when I build a raidz of 4 >> "marvell" disks when compared to a raidz on the Areca controller. >> >> The essential difference is that on the basic controller the disks are >> recognised as being 4K disks with quirks. On the Areca controller there >> is no quirk that matches... >> >> Is this a typical problem on the Areca controller? Or does this not work? > > The Marvell controller is recognized as a SATA controller, so the drives > attached to it are treated as SATA drives, using the ATACAM stack: > > ada0: <WDC WD40EFRX-68WT0N0 80.00A80> ATA-9 SATA 3.x device > > > The drive in question is listed along with the SATA quirks: > > > sys/cam/ata/ata_da.c :: ada_quirk_table[] > { > /* WDC Caviar Green Advanced Format (4k) drives */ > { T_DIRECT, SIP_MEDIA_FIXED, "*", "WDC WD????RX*", "*" }, > /*quirks*/ADA_Q_4K > }, > > > The Areca controller appears to be acting as a *SAS* controller, so the > drives attached to it are treated as SAS drives, using the SCSICAM stack: > > da0: <WDC WD40EFRX-68WT0N0 R001> Fixed Direct Access SCSI-5 device > > While the SCSICAM stack knows about quirks for SCSI/SAS drives, it doesn't > know anything about quirks for SATA drives. > > There may be an argument there to create a unified quirk list, but that's > a larger discussion. That is probably is correct summation of my question... I more or less expected things not to be different in such extend. --WjW
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?531F3533.7000407>