Why is disk speed so slow on Dedicated CPU instance?
I made a 32GB dedicated CPU instance. Upon running:
dd if=/dev/zero of=/tmp/test1.img bs=1G count=1 oflag=dsync
I received the output:
1073741824 bytes (1.1 GB) copied, 132.879 s, 8.1 MB/s
2nd run was better, but not great:
1073741824 bytes (1.1 GB) copied, 9.56482 s, 112 MB/s
Subsequent runs when pretty fast, at approximately 720 MB/s, but not always. At the 6th run I got:
1073741824 bytes (1.1 GB) copied, 87.1838 s, 12.3 MB/s
hdparm -t
also showed inconsistent numbers, once being 12 MB/sec, once being 322 MB/sec, once being 565 MB/sec, all of which are slower than my regular standard instance, which often gives 800 - 1100 MB/sec, and that's while other things are in use.
I tried this on two instances and had similar results on both.
Any idea why this is happening? Do the servers need some sort of "warm-up"? to go at full speed?
1 Reply
I'm not quite sure why your Linode's seeing inconsistent disk speeds, but I'd be happy to look into it with you. Could you open up a ticket for your Linode? Once you do, myself or another member of our Support team will check things from our end to make sure everything's running smoothly.