Date: Thu, 10 May 2007 23:25:04 +0200 From: Peter Schuller <peter.schuller@infidyne.com> To: Peter Schuller <peter.schuller@infidyne.com>, freebsd-questions@freebsd.org Subject: Re: Stable many-port SATA controller recommendations Message-ID: <46438DB0.5050107@infidyne.com> In-Reply-To: <20070510203217.GA88574@voi.aagh.net> References: <46435C68.9070505@infidyne.com> <20070510203217.GA88574@voi.aagh.net>
next in thread | previous in thread | raw e-mail | index | archive | help
This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enig6561AC25DCC9D98090704C2F Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable > Promise are supposed to be one of the better makes when it comes to > documentation and open source support,=20 Yes, that was my impression prior to purchasing the two TX4:s. > what problems are you seeing? I can easily (dd if=3D/dev/zero of=3Dtestfile bs=3D$((1024*1024)) count=3D= 500) trigger within a few seconds: WARNING - WRITE_DMA48 UDMA ICRC error (retrying request) LBA=3D<a number>= Followed perhaps 30 seconds later by the usual flooding of g_vfs_done() errors and a panic resulting from the inability to flush dirty buffers. If anyone is interested I can provide actual details, but there's nothing really interesting to see as far as I can tell. I have tried several different SATA cables (since in other contexts with other controllers these messages were supposedly often indicative of bad cabling). The TX4, or at least this particular type, is specifically a PDC40718. I did talk to someone on ##freebsd who indicated he had not had any trouble with this particular chipset, so I presume it is somehow being triggered by the environment. It's an Epox EP-8KMM5I motherboard. > It's supported, and even better, the Marvell is supposedly now 64bit > clean with ata(4) in HEAD. When I tried it a year or so ago it would, > like any other ATA controller, explode with more than a couple of disks= > and >=3D4GB RAM because of problems with the way bounce buffers are > allocated; I didn't see any problems aside from that, but I didn't test= > it very much. Ok. 64bit/4 GB is not an immediate problem for me, though it will be relevant in the future. But nice to hear someone had it working stably. > When ata(4) proved itself utterly useless in environments with decent > amounts of memory, I got an 8 port LSI MegaRAID, since amr(4) had just > been given an overhaul to make it MPSAFE and 64bit clean; I've not had > any real problems with it (aside the occasional fight with the BIOS, > which is par for the course with RAID controllers), but I'd still prefe= r > a real SATA controller so things are a little more predictable and > flexible (SMART support, software RAID, disks exposed to the actual OS > and not some dodgy half-assed ASIC..). Yes. Given that I do not intend to utilize the hardware RAID support, having to go for a RAID card is to me just a downside, even disregarding the increase in cost. Thank you for the input! --=20 / Peter Schuller PGP userID: 0xE9758B7D or 'Peter Schuller <peter.schuller@infidyne.com>' Key retrieval: Send an E-Mail to getpgpkey@scode.org E-Mail: peter.schuller@infidyne.com Web: http://www.scode.org --------------enig6561AC25DCC9D98090704C2F Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.3 (FreeBSD) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGQ426DNor2+l1i30RCFs4AJ9TNq/TIdHNThxaJK7gBemUXCOjJgCfZDnU 6BIJWzsNe4gBWFgztNo0HvQ= =RhKZ -----END PGP SIGNATURE----- --------------enig6561AC25DCC9D98090704C2F--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46438DB0.5050107>