4

I've recently (for about a month or so) been having issues with one of my secondary hard drives. When accessing data it, sometimes it will hang for a few seconds, causing the program doing the reading to also lag. For example, in video games being run off the drive, about every 30 seconds the game will hang for about 4-5 seconds, and in Visual Studio saving a file often causes the whole IDE to hang for about 5-10 seconds.

I have two identical drives (Hitachi 7K3000 - 2TB), and only one of them is displaying this issue. I ran HD Tune on both to compare their read speeds:

The faulty drive: http://i39.tinypic.com/8x0dox.png

The normal drive: http://i41.tinypic.com/o0arft.png

Note how in the first image it took some time at the beginning to start.

I also ran the "Extra tests" in HD Tune:

http://i40.tinypic.com/keygcg.png - Faulty drive (note the ridiculous initial random seek time)

http://i43.tinypic.com/28vbtd5.png - Normal drive

How might I fix this issue? Maybe copy the data to the currently normal hard drive (which doesn't have anything useful on it) and reformat the faulty one? The drive is still under warranty, so maybe just RMA it? Other suggestions? My boot drive is a separate SSD so my options are pretty flexible.

8
  • Also as a secondary thing, does anyone know why there is such a discrepancy in the burst rate? The cache performs totally differently between the two seemingly identical drives..
    – jli
    Jan 31, 2012 at 17:14
  • Interesting how your min write, sometimes the most important, is taking a dive. The drive has data on it? the drive is in active use somehow ? If that is true how would that effect the testing? especially when testing is less able to do completly direct communications to the drive anymore. It is hard enough to test a blank drive , but a drive with data on it , and in a system :-) well so many factors involved.
    – Psycogeek
    Jan 31, 2012 at 17:42
  • @Psycogeek the drive has data on it, but was not being used during the test. I could try formatting it and retesting, that might be better.
    – jli
    Jan 31, 2012 at 18:36
  • I was wondering More that something else might be going on, other than the test. More than it having data on it.
    – Psycogeek
    Jan 31, 2012 at 19:00
  • 1
    So I just unmounted both drives completely and reran the tests, and there was no change.
    – jli
    Jan 31, 2012 at 19:17

3 Answers 3

3

Have you considered the possibility of the connections on the controller? Like, swap which sata cable is connected to each, and which port on Mobo or Sata Controller? It is rare, but it is still minutely possible the issue exists somewhere between the HDD and the mobo, and not the drive itself.

1
  • although this is more unlikely, its alot less of a hassle to address if it is the issue, and could take minimal time to test.
    – DaBaer
    Jan 31, 2012 at 17:45
3

I'd return the drive that's behaving oddly immediately. The two things I can think of that would cause intermittent pauses in the data stream are faulty power management firmware suddenly spinning the drive down, and soft read errors causing multiple retries to read a block that's gone bad. In both cases, the drive isn't going to get any better no matter what you do.

0
1

As DaBaer indicated, re-seat everything. Even if the drive itself is the problem, Hitachi support will likely have you do it regardless as it will potentially prevent the necessity of a drive RMA. I worked in hardware support for years before I became a developer and was personally amazed at number of times a simple connector re-seat solved performance / functionality issues. Also it should be quick to test with SATA connectors.

Beyond that backup any important data into a compressed archive and write it to both drives so you're in no danger of losing anything.

Hitachi support likely has a preferred low-level-format utility they can direct you to for your specific HDD model. After you run it re-run those HDD tests and see if an RMA needs to be pursued further.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.