Why VMware Virtual disk so slow (da0: 3.300MB/s transfers) ?

Dongsheng Song dongsheng.song at gmail.com
Tue Aug 20 23:52:07 PDT 2013


On Wed, Aug 21, 2013 at 5:12 AM, Bryan Vyhmeister <bryan at bsdjournal.net> wrote:
> On Aug 20, 2013, at 14:05, Justin Sherrill <justin at shiningsilence.com> wrote:
>
>> On Mon, Aug 19, 2013 at 10:15 PM, Dongsheng Song
>> <dongsheng.song at gmail.com> wrote:
>>
>>> 4156467 bytes/sec = 3.964 MB/s
>>
>> I don't have a good answer for you other than that it's not endemic to
>> DragonFly - this is what I got under VMWare Workstation:
>>
>> 1073741824 bytes transferred in 24.643408 secs (43571158 bytes/sec)
>>
>> - Which is about 10 times your rate.  I get the same 3.300MB/sec
>> transfers notice in my dmesg, so it may be something with the
>> underlying disk.
>
> Perhaps this has already been addressed but normal hard disks under VMware ESXi are dreadfully slow unless you have a battery-backed RAID controller. I've done local testing with SSDs instead which pretty much alleviates the speed issues but gives no redundancy.
>
> Bryan

I don't think this is a battery issue, since Linux guest in the same
ESXi host give me very good results:

root at CRM-SYSLOG:~# dd if=/dev/zero bs=8M count=128 of=/tmp/1G.img
128+0 records in
128+0 records out
1073741824 bytes (1.1 GB) copied, 2.10386 s, 510 MB/s

root at CRM-SYSLOG:~# dd if=/dev/zero bs=8M count=128 of=/tmp/1G.img
128+0 records in
128+0 records out
1073741824 bytes (1.1 GB) copied, 7.08881 s, 151 MB/s

root at CRM-SYSLOG:~# dd if=/dev/zero bs=8M count=128 of=/tmp/1G.img
128+0 records in
128+0 records out
1073741824 bytes (1.1 GB) copied, 6.85367 s, 157 MB/s




More information about the Users mailing list