locked
Very Slow Throughput on WHS Drives in Storage Pool (but not other drives) RRS feed

  • Question

  • My WHS installation worked great for months, but for the last month or so performance has been awful.  I checked using HDTune, which shows throughput of 2 MB/sec on my system drive and the 2 other drives in the storage pool.  (The drives are all SATA, 2 750 GB drives and a 640 GB drive). 

    In contrast, another 500 gig SATA drive in the server that I have not added to the server pool shows 90 MB/sec throughput, and even a USB-attached drive configured as a backup drive shows 35 MB/sec. throughput, so I don't think there's anything wrong with the motherboard or drivers.

    All drives show as healthy in the console, but running chkdsk /f/r (after stopping the backup service) hangs at about 10% in step 4.

    I am using PowerPack 3, but had the same problem with PowerPack 2.

    Any ideas appreciated.  Thanks.

    Tuesday, September 8, 2009 6:18 PM

Answers

  • No, one bad drive should normally result in a single channel running in PIO mode. Still, go through and check all cabling (I would just replace the SATA cables, to be honest; even decent quality cables are cheap), make sure power connectors are properly seated, etc. You can't reset to UDMA mode manually, unfortunately. However, you can uninstall the device and Windows will normally redetect it on next boot and reset the transfer mode for you. See this support article (the Workaround section) for more information. If it was some transient issue, probably you'll be fine, but if there's a hardware issue somewhere, you might need to do some more work to identify the failing component(s).

    By the way, SMART turns out not to be a good predictor of failure in hard drives. Google did some research on this a couple of years ago (Google being one of the largest consumers of hard drives in the world). What they found was that certain SMART indicators do suggest that a drive is having trouble, but the lack of SMART errors doesn't indicate that a drive isn't going to fail tomorrow. And SMART doesn't tell you anything at all about other components of your storage subsystem, just the drives themselves.

    I'm not on the WHS team, I just post a lot. :)
    • Marked as answer by gleklen Thursday, September 10, 2009 1:35 PM
    Wednesday, September 9, 2009 3:50 PM
    Moderator

All replies

  • When chkdsk hangs, it's indicative of a serious problem with your storage subsystem. So first, can you log in to your server and make sure that none of your drives are running in PIO mode? That's another indicator of drive, controller, or cable failure.
    I'm not on the WHS team, I just post a lot. :)
    Tuesday, September 8, 2009 9:48 PM
    Moderator
  • Thanks.  Two of the channels show that they're running in PIO mode.  I'll try swapping cables and see if that works.  Otherwise, would one bad drive result in the two channels showing PIO mode?  Is there any good way to figure out which drive it is?  Nothing shows as a problem in the SMART analysis in Speedfan.

    Thanks again.
    • Marked as answer by gleklen Thursday, September 10, 2009 1:35 PM
    • Unmarked as answer by gleklen Thursday, September 10, 2009 1:35 PM
    Wednesday, September 9, 2009 2:44 AM
  • No, one bad drive should normally result in a single channel running in PIO mode. Still, go through and check all cabling (I would just replace the SATA cables, to be honest; even decent quality cables are cheap), make sure power connectors are properly seated, etc. You can't reset to UDMA mode manually, unfortunately. However, you can uninstall the device and Windows will normally redetect it on next boot and reset the transfer mode for you. See this support article (the Workaround section) for more information. If it was some transient issue, probably you'll be fine, but if there's a hardware issue somewhere, you might need to do some more work to identify the failing component(s).

    By the way, SMART turns out not to be a good predictor of failure in hard drives. Google did some research on this a couple of years ago (Google being one of the largest consumers of hard drives in the world). What they found was that certain SMART indicators do suggest that a drive is having trouble, but the lack of SMART errors doesn't indicate that a drive isn't going to fail tomorrow. And SMART doesn't tell you anything at all about other components of your storage subsystem, just the drives themselves.

    I'm not on the WHS team, I just post a lot. :)
    • Marked as answer by gleklen Thursday, September 10, 2009 1:35 PM
    Wednesday, September 9, 2009 3:50 PM
    Moderator
  • Uninstalling and the devices seems to have fixed it.  I'll see if it fails again.

    Many thanks again for your help.
    Thursday, September 10, 2009 1:35 PM
  • I see a similar problem on my WHS setup. Drives sometimes fall back to PIO mode if drive spindown is enabled.

    I'm speculating that my combination small power supply (Chenbro ES34069 case 120W power supply) and older drives with high power drain at spinup is a problem. Once I have phased out the last two old drives for Caviar Green drives I'll try spindown again.
    Tuesday, September 22, 2009 2:28 PM