Rackspace to RDS using DMS (Postgres 9.2)

On 08/31/2016 03:41 PM, Patrick B wrote:

Hi guys, I posted this question on the ADMIN list but will post here as
well so more people can comment...
/messages/by-id/CAJNY3it_AfxJhmwMHtpiAbHG47GS5rJOAUgfHw+Gm5OXCbUm7w@mail.gmail.com

I've got a 2.3TB Database running at Rackspace... We'll be migrating it
to RDS PostgreSQL 9.5 very soon...

We already have an EC2 Instance at Amazon running PostgreSQL 9.2 as
streaming replication from Rackspace.

I'll have to upgrade the version of Postgres on that instance before
start using DMS service.

*Question:*
Has anybody ever used that service? I'm just trying to find out how much
time it will take to perform the migration...

It is 2.3TB, it is going to take a long time no matter what service you
are running.

No, I have not used DMS. Frankly, with all respect to AWS/RDS the idea
of running a 2.3TB instance that will get any level of performance
sounds ridiculously expensive.

Sincerely,

JD

Thanks!
Patrick

--
Command Prompt, Inc. http://the.postgres.company/
+1-503-667-4564
PostgreSQL Centered full stack support, consulting and development.
Everyone appreciates your honesty, until you are honest with them.
Unless otherwise stated, opinions are my own.

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

From: Joshua D. Drake Sent: Thursday, September 01, 2016 9:29 AM
On 08/31/2016 03:41 PM, Patrick B wrote:

Hi guys, I posted this question on the ADMIN list but will post here

as well so more people can comment...

</messages/by-id/CAJNY3it_AfxJhmwMHtpiAbHG47GS5rJ&gt; /messages/by-id/CAJNY3it_AfxJhmwMHtpiAbHG47GS5rJ

OAUgfHw%2BGm5OXCbUm7w%40mail.gmail.com

I've got a 2.3TB Database running at Rackspace... We'll be migrating

it to RDS PostgreSQL 9.5 very soon...

We already have an EC2 Instance at Amazon running PostgreSQL 9.2 as

streaming replication from Rackspace.

I'll have to upgrade the version of Postgres on that instance before

start using DMS service.

*Question:*

Has anybody ever used that service? I'm just trying to find out how

much time it will take to perform the migration...

It is 2.3TB, it is going to take a long time no matter what service you are running.

No, I have not used DMS. Frankly, with all respect to AWS/RDS the idea of running a 2.3TB instance that will get any level of performance sounds ridiculously expensive.

Sincerely,

JD

I currently have an EC2 instance in AWS – an m4.xlarge (4 cores, 16gb, 3tb SSDs) and it’s pretty cheap, about $620/mo ($210/mo for the compute, $410 for the storage). The performance of this setup rivals in-house Cisco UCS server that we are demoing that costs ~$100k, as long as our batch sizes don’t exceed available memory – that’s where the larger Cisco pulls ahead. The $620/mo is the on-demand price, btw…the reserved price is much lower.

$100k/ $620 = 161 months of operation before cost parity.

Mike S

It is 2.3TB, it is going to take a long time no matter what service you
are running.

I know that Joshua! I just wanna get an idea from someone who has used DMS
service...

No, I have not used DMS. Frankly, with all respect to AWS/RDS the idea of
running a 2.3TB instance that will get any level of performance sounds
ridiculously expensive.

A business that has a 2.3TB database large, is not worried about costs...
the system has been running fine since 2009 and now it's time to change the
infrastructure to something more easily to make changes when needed.... We
also have lots of customers in Australia and USA, so the Viability at
Amazon will help with the latency for sure.

Sincerely,

JD

I currently have an EC2 instance in AWS – an m4.xlarge (4 cores, 16gb, 3tb
SSDs) and it’s pretty cheap, about $620/mo ($210/mo for the compute, $410
for the storage). The performance of this setup rivals in-house Cisco UCS
server that we are demoing that costs ~$100k, as long as our batch sizes
don’t exceed available memory – that’s where the larger Cisco pulls ahead.
The $620/mo is the on-demand price, btw…the reserved price is much lower.

$100k/ $620 = 161 months of operation before cost parity.

Mike S

We'll be using RDS, not an EC2 Instance.

On Thu, Sep 1, 2016 at 1:44 PM, Mike Sofen <msofen@runbox.com> wrote:

$100k/ $620 = 161 months of operation before cost parity.

You ought to maybe compare what you can buy from an outfit like
Silicon Mechanics for ~$15k. I suspect you could get 16-core, 256GB
RAM, and several TB of disk. And then your parity would be much
shorter and you'd have higher performance.

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general