Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a few servers. Each has 9 backplanes, and each backplane has 5 hard drives. Once in a while, one of the cables connecting a backplane to the motherboard (actually a pci card on the motherboard) fails, and the system starts throwing errors, and keeps resetting the SATA link. But this happens only when there is a lot of load on a particular link. Replacing the cable sets everything alright.

My question is, is there a tool (for Linux) that can stress test all the SATA links, so that I can know which cable is faulty before deploying a server for production? Something that writes excessively to all the hard drives? (Note: hard drives are raided in groups of 15, and then a logical volume is made out of the 3 raid sets. When I try writing a lot of data to the volume, the OS usually only writes to one raid array.)

share|improve this question

badblocks will write to every sector of a drive (random data or zeros, at your choosing) and then reread it to verify.

I'm unclear on how a RAID setup would prevent writing completely to all drives unless your RAID containers are smaller than the capacity of your drives (I'm not terribly experienced with RAID so I may be wrong). You'd need to undo the RAID configuration and let them be directly addressable individually by Linux to test with badblocks fully.

share|improve this answer

Hey Skipper it sounds like you have a BackBlaze Pod <=3.0 and you are trying to check the SATA link to the port multiplier (and possibly also from the multiplier to the attached drives.

share|improve this answer
How it answers the question?could u elaborate your answer? – BlueBerry - Vignesh4303 Oct 4 '13 at 4:01
It does not answer the question it may clarify it somewhat. I did a minute search on google and found nothing immediately related to what I think the question is. which means one of the following nobody knows anything useful, somebody knows but is not posting about it or most likely I am too thick to ask the right question. – JamesB192 Oct 4 '13 at 15:12
I remember seeing the blog entry I linked above mentioned issues with flaky connectors on SATA cables and I thought they might liaise. My wrong headed idea that the SATA is essentially a special type of computer network and what would probably work best is something almost but not completely unlike the hash column of a rainbow table fed into ping -p. – JamesB192 Oct 4 '13 at 15:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .