Driven by the current price difference between SATA and SAS disks on one side and the potentially bad behaviour of SATA disks in bigger storage arrays on the other side, I have found so-called SATA-to-SAS interposer cards.
Advertised as "seamlessly adding SAS capabilities to existing SATA disk drives", I wonder if anyone here has had some experience with these or similar products. The major benefits I can identify are the increased cable voltage (if all drives are SAS connected), the ability to power-cycle the drive and multipath (if desired). Obviously the SATA drive will still have to be RAID edition.
The question is: Do these cards indeed increase the overall reliability of a storage system, or will failing SATA disks cause trouble nevertheless?
Edit: I'm not asking for hypothetical answers, only actual experience please.
I'm well aware that the typical 10k SAS drive is more reliable (and better performing) than 7200 SATA drives. But how does a nearline SAS, which is phyiscally the same disk as its SATA counterpart, compare to the SATA version with interposer?
My two cents: If you are concerned about the edge case failures that may occur with SATA hardware (specifically, lousy SATA controllers), spend the money on real SAS disks.
These cards do what they say on the tin: They translate SAS (SCSI) commands to SATA commands, and even implement a few themselves (like power- and spinup-control).
They do nothing else (the drives are still SATA, their performance characteristics are unchanged, their reliability is unchanged, etc.), and they add a new layer of complexity into your environment (They are hardware + software, either of which could fail, have a defect, etc.) -- From my point of view you are increasing your net chance of a failure versus just buying an appropriate SAS drive.
There is one main reason I would use them - if I was maxing SATA and SAS drives on the same backplane. As they use different signalling levels, it's not recommended to use both on the same backplane.
As these should convert the SATA interface to SAS and vice versa, these mitigate this potential issue.
The only other reason I can think of is if I had some drive trays/caddies for a server that were intended for SAS drives only and I wished to fit SATA drives to these. In this case the interposer should compensate for the different screw positions.
Using interposer does give you a so called native SAS experience to the drive but the trade off of that being the error handling and recovery mechanism which is also delegated to these interposer devices. Although there is the T10 written document for SCSI to ATA translation (SAT) but the finer details are left to the implementer. A case in point is as follows - SATA does not have a notion of abort command which is used in SAS domain for recovering a command. So when the SAS interposer has a command which needs to be aborted by the host it will translate the abort into a SATA equivalent of device soft reset an inadvertent effect of force evicting all active commands with the SATA drive and hence cause latency and other subtle failures (I can fill the details if necessary). You could say we should avoid host issuing abort and this problem would not occur. Sure enough but the reverse case, in case the drive encounters an error which leads to an over/under run the interposer has nothing else but to cause a device reset to clear that condition essentially leading to the same effect as abort. In the later example, host has no control and is nature of system.
In some situations you maybe better off no to use the interposer and use native SATA command set. Most SAS controllers support both SAS and SATA attachment and allow a mix as well. But if the requirement is dual port access to SATA drives you are locked into getting an interposer. Alternatively there is a class of drives which are coming along called the FAT SAS drives (fat implies the capacity and not the physical form factor) which are a viable alternative albeit the drive reliability is definitely less than that of the $$ sas drives.
I think the question is already answered but I'm going to add something else, for my experience, even is a device that usually doesn't fails, is a new point of failure, imaging a situation where you find a hard drive failure, normally is going to be the hard drive, but before changing the drive, be sure the interposer is working properly.
So of course there is some trade-off, but I'm not sure the current answers provide any positive use cases. Lets say the biggest benefit of the interposer is the multi-path, which is removing a fairly large single point of failure from the system. Is it worth it for that feature alone?
The comparison can't be against buying a SAS drive, because that is spending 2x the price vs $40 on an interposer. If you can trade an increased chance of losing a single drive (interposer) for adding a fully redundant active-active secondary host, that's got to be worth $40 / drive, right?
You said no theoretical advice, and I think that's a good policy. I am deploying a subset of enclosures with SATA-SSD with SAS interprosers, specifically for the purpose of having two direct-attached hosts. I have heard that there are companies that build the enclosures exactly this way by design. The problem, I think, is even sourcing the interposers, and practically, fitting them in the enclosure. But maybe it's fair to say, they should be used where-ever practical.