1

I'm a front end guy, and I've been hiring a back end developer overseas to develop my API for me.

We've been testing and developing at AWS on their EC2 on a t2.micro instance. My backend guy has quoted me an absolute fortune to upgrade to a t2.medium plan (around 15 hours), but far less (1 hour) to upgrade to a t2.large plan.

Why is upgrading to a medium instance far easier than upgrading to a large instance?

EDIT: Developer has now changed his mind and is stating that all the T2 instance types cannot be upgraded to without a full reinstall. Apparently we can upgrade to m3.large, but no T2 instances. Note the picture below:

enter image description here

Fizzix
  • 23,679
  • 38
  • 110
  • 176
  • 1
    No obvious reason why this should be the case. Get him to elaborate. – John Greenall Feb 09 '16 at 12:10
  • @JohnGreenall - All he said was that our current plan does not support HVM – Fizzix Feb 09 '16 at 12:11
  • Apparently upgrading to a large plan is seamless, but upgrading to a medium plan requires a full rebuild (installing PHP, MySQL, etc) of the server. – Fizzix Feb 09 '16 at 12:13
  • At this moment I would take control of your destiny and do a little more reading about EC2. In theory, this should be as simple as creating an AMI from your existing instance, launching a larger instance from that AMI then updating your Apache/ Nginx/ NodeJS settings to take -advantage of more threads/ memory as per Benoit's advice. – John Greenall Feb 09 '16 at 13:40
  • 2
    I can do that for you for free :) Seriously, 15 hours is just too much. – helloV Feb 09 '16 at 15:19
  • There's a lot of bad info coming your way. "Upgrading" (or downgrading) in instance within the t2 class takes 2 minutes, max; additionally, the t2.micro **only** supports HVM. Are you sure you're not on a **t1**.micro (which is PV, not HVM)? (Even still, there would be no difference between t2.medium and t2.large -- the entire t2 family from nano to large is HVM only). – Michael - sqlbot Feb 09 '16 at 21:21
  • @JohnGreenall - It has been in the back of my mind to take over the backend stuff since I have a lot of experience with PHP, MySQL and Apache. Just not too great with any command line stuff. – Fizzix Feb 09 '16 at 21:30
  • the latest screen shot shows that a m3.large can't go to a t2.anything because the t2 instance types do not support hvm and the m3 do. Which isn't apposite if you are going from t2.micro to a larger t2 type – Vorsprung Feb 09 '16 at 21:34
  • personally I would pay the guy for a t2.micro and get someone reliable to upsize the t2.micro to a required size – Vorsprung Feb 09 '16 at 21:35
  • 2
    It sounds like you currently must have a t1.micro, then, not a t2. There's absolutely no reason you couldn't go to a different t2, if that's what you already have... but with t1.micro, yes, you'd need to rebuild for HVM to go to t2... t2 => HVM only, t1 = PV only, m3 => HVM or PV. – Michael - sqlbot Feb 09 '16 at 21:54
  • Just go to the EC2 dashboard, stop the instance, change the type, and restart it. – NeilG May 21 '23 at 07:08

2 Answers2

2

If you look at this page

http://www.ec2instances.info/?selected=t2.small,t2.medium,t2.large

The cost per hour is more for a small, medium and large

On our systems we routinely shut them down and restart them on a different instance type. This typically takes a few minutes and no software changes. Altering the application memory use might be required I suppose but this is a trivial change

If the extra charges include the hosting for days/months then the difference in pricing makes sense

Using "on demand" pricing for the US-East zone there is a $62 difference in price between a t2.small and a t2.large per month

Vorsprung
  • 32,923
  • 5
  • 39
  • 63
  • "shut them down and restart them on a different instance type" Should I create an image? Invite you to answer this question, [How do I upgrade from t2.micro to t2.small at the moment of expiration?](http://stackoverflow.com/questions/43690262/how-do-i-upgrade-from-t2-micro-to-t2-small-at-the-moment-of-expiration). – SparkAndShine Apr 29 '17 at 00:10
  • @SparkAndShine The other answers to that question pretty much cover it. There is no need to make an image – Vorsprung May 01 '17 at 08:11
1

As a daily user of AWS, I can say that the only operations when upgrading within the same generation of instances (here t2) is to resize the VM memory and threading configuration.

1 hour is reasonable as your backend developer should update its deployment script and redeploy the infrastructure, but 15 hours sounds like a typo to me.

Benoît Sauvère
  • 701
  • 7
  • 23
  • I can confirm it's not a typo, unfortunately. Have asked him, and he quoted the amount in cash and it's definitely 15 hours worth, if not more. – Fizzix Feb 09 '16 at 12:15
  • And the 1 hour for t2.large is not ? – Benoît Sauvère Feb 09 '16 at 12:16
  • Correct. Can confirm that it was quoted at 1 hour for large and 15 for medium. All he said that it was sometime to do with our current plan not supporting HVM or something like that? – Fizzix Feb 09 '16 at 12:27