badblocks extremely slow

Ferry
2013-03-05
2013-03-05
  • Ferry

    Ferry - 2013-03-05

    Hi there,

    we just got a new (used) server with 12 1TB SAS drives. These things do about 140MB/s (outter ring) sequential with dd (dd if=/dev/zero of=/dev/sda bs=1M) after the cache is full. Both iostat (-x 2) and dd's output (kill -USR1) verify this.

    Badblocks does around 7.8MB/s consistently on all 12 drives. iostat will show them fully loaded (utilization is >= 98% consistently).

    I've never seen this. Usually badblocks speed more or less equals dd's speed.

    The only thing I can image is that it's some how related to the USB stick I'm running ubuntu off. It's not even really the USB stick, it's the port, it's only USB 1.1 (why I don't know, the chipset (lshw) clearly shows it supports USB 2.0 - but why on earth would one wire that right? USB 1.1 oughtta be enough for everybody!). Find that weird tho', should cause much disk I/O for a consistent pattern write (badblocks -svw /dev/sdX).

    Any ideas? Quite convinced it's not really an issue (well besides badblocks taking ages & ages) but it has me worried nonetheless.

     
  • Ferry

    Ferry - 2013-03-05

    Err should read shouldn't. Badblocks shouldn't cause much disk I/O for consistent pattern. To be more exact, to the 'root' volume / the USB stick badblocks is running from, it obviously causes a ton of disk I/O on the disk it's testing :).

     
  • Ferry

    Ferry - 2013-03-05

    Would have sworn I already posted it works fine (same speed as dd sequentially) with the -B option, but don't see it here anymore.

    Anyways, it works fine with the -B option, should I be worried about the Direct I/O being so slow though?

     

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks