Roman Mamedov put forth on 4/14/2011 5:54 AM: > I suggest that you avoid Silicon Image 3132, they have data corruption issue > (on some board designs?), triggered or amplified by transferring via both > ports at the same time, at full speed. Stating this opinion based on a very limited number of such reports is irresponsible. If the issue is a bad PCB from a couple of vendors, state the vendors and the card model and rev. DO NOT paint the SATA chip itself as BAD and something to avoid. Case in point: http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/ BackBlaze uses three Sil3132 based 2 port Syba PCIe x1 SATAII cards in each storage 45 drive storage pod server, and one 4 port Addonics Sil3124 PCI card. Odd balance WRT bandwidth, but if you read the doc it'll be clear why they did this. They've had HUNDREDS of these Syba and Addonics Silicon Image chip cards in production for multiple years now and report nothing but great success. They use 9 SATAII backplanes per chassis each with a 1:5 Silicon Image 3726 PMP, for a total of 45 drives per chassis. They create 3 x 15 drive mdadm RAID6 arrays per chassis, as the drives are physically laid out in 3 rows of 15. These systems are greater in scale and complexity than anything I've seen discussed on this list in the year+ I've been subbed. Again, the backbone of their 45 drive pod is the Sil3132 chip, each 3132 managing 10 drives, 5 each connected to BOTH card ports via PMPs, 30 drives total per chassis controlled by the Sil3132 chips. The Addonics Sil3124 card manages 15 drives across 3 ports and 3 PMPs. Silicon Image has the most compatible, tested, and field proven reliable SATA chips on the planet, ditto its PMP chips. BackBlaze is but one example of this. Also, Silicon Image PMP chips ship in more products than all other PMPs combined. Again, there is nothing wrong with the Silicon Image chips. They have a very good reputation. There is apparently a problem with some Chinese really cheap badly designed/manufactured Silicon Image based cards going into the Russian market. From the comments I read it appears the "sample set" of bad cards consisted of less than a dozen. This could have simply been a bad batch of cards. From the comments, this actually seems like the most likely scenario. -- Stan -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
