I'm running a Zope web application that obtains most of its data
from a PostgreSQL 8.1 database in a virtual machine. I'm able to
adjust the memory of this machine according to reasonable values
and can choose between one or two (emulated) processors. The
question is: How can I find an optimal relation between the
virtual hardware parameters and PostgreSQL performance. I guess
it makes no sense to blindly increase virtual memory without
adjusting PostgreSQL configuration. Are there any experiences
about reasonable performance increasing strategies? Are there any
special things to regard in a VM?
I am responsible for some active PostgreSQL databases within virtual
machines and on "plain metal"; all using Windows.
There are spurious strange performance issues on those PostreSQL
databases within the virtual machines. I can not pin them down (yet),
and am not really able to blame them on vmware; but just want to
recommend to be very very carefull about PostgreSQL in vmware and
please communcate your findings.
> from a PostgreSQL 8.1 database in a virtual machine. I'm able to
> adjust the memory of this machine according to reasonable values
> and can choose between one or two (emulated) processors. The
> question is: How can I find an optimal relation between the
> virtual hardware parameters and PostgreSQL performance. I guess
You did not specify the OS in your VM. I can report my experiences
with W2k3 inside the VM:
contrary to usual recommendations, smaller shared_buffers yielded
Growing effective_cache_size yielded the best results for me.
GHUM Harald Massa
persuadere et programmare
Harald Armin Massa
Python: the only language with more web frameworks than keywords.
Andreas Tille wrote:
> Are there any experiences
> about reasonable performance increasing strategies? Are there any
> special things to regard in a VM?
Not directly about Postgresql, but I'm seeing evidence that upgrading
from vmware 2.5.3 to 3.0.1 seems to have solved disk access
performance issues (measured with simple dd runs).
With vmware 2.5.3 + RedHat Enterprise 4.0 I measured a sequential
read performance on 1-2 Gb files of less than 10 Mbytes/sec on a
IBM FastT600 SAN volume partition.
After the upgrade to 3.0 I had feedback from sysadmins that issue
was solved, but I didn't have the opportunity to repeat the read
Your message is accepted. The reference number is .