News for hackers. Stuff that matters.
Re:Disk IO and throughput benchmarks on Amazon's EC2 by spudlyo on Wednesday December 09, 2009 @05:02PM (#986701)
It's worth noting that you can speed up write operations on EBS volumes by first initializing them by writing data to every block on the volume. Subsequent writes will be faster due to the way Amazon virtualizes disks.

http://docs.amazonwebservices.com/AWSEC2/2007-08-29/Develope...

Re:Disk IO and throughput benchmarks on Amazon's EC2 by aolnerd on Wednesday December 09, 2009 @06:00PM (#986822)
Those docs refer to ephemeral storage. I haven't seen any claims by Amazon or test results that indicate it is also true for EBS volumes. I have been meaning to test this myself.
Re:Disk IO and throughput benchmarks on Amazon's EC2 by spudlyo on Wednesday December 09, 2009 @08:47PM (#987115)

  gracie:~# time dd if=/dev/zero of=/dev/sdl bs=1024k count=10240
  10240+0 records in
  10240+0 records out
  10737418240 bytes (11 GB) copied, 344.745 seconds, 31.1 MB/s

  real    5m44.748s
  user    0m0.025s
  sys     0m10.152s

  gracie:~# time dd if=/dev/zero of=/dev/sdl bs=1024k count=10240
  10240+0 records in
  10240+0 records out
  10737418240 bytes (11 GB) copied, 217.106 seconds, 49.5 MB/s

  real    3m37.125s
  user    0m0.023s
  sys     0m10.268s

  gracie:~# time dd if=/dev/zero of=/dev/sdl bs=1024k   count=10240
  10240+0 records in
  10240+0 records out
  10737418240 bytes (11 GB) copied, 247.319 seconds, 43.4 MB/s

  real    4m7.323s
  user    0m0.036s
  sys     0m9.794s
EDIT: I ran it a third time just to make sure the results held. I'd still take this with a grain of salt, in my experience EBS i/o results are pretty inconsistent.
Re:Disk IO and throughput benchmarks on Amazon's EC2 by 3ijsflksflk on Wednesday December 09, 2009 @08:56PM (#987127)
data rules