I want to run 10 instances of Massplanner 24/7

Kansept

BANNED
Joined
Jul 14, 2016
Messages
138
Reaction score
22
What's the best way to do this?

10 VPS

A dedicated server with partitions

Or should I just buy 1 beefy VPS? MP-Admin said I can get threaded connections up to 200 or 300 if my system can handle it.

ANY information about amping up Massplanner would be appreciated!
 
I would recommend multiple VPS servers on a provider with a easy to use interface, that way scaling up would be alot easier.

Oh, why not a dedicated instance from Vultr with addon IP's? That could easilly run 10 instances of Massplanner. What about OS?
 
What's the best way to do this?

10 VPS

A dedicated server with partitions

Or should I just buy 1 beefy VPS? MP-Admin said I can get threaded connections up to 200 or 300 if my system can handle it.

ANY information about amping up Massplanner would be appreciated!
Whelp, someone's rolling in the dough :p

The best way to have 10 concurrent instances of massplanner open at one time is to figure out the maximum number of virtual machines you can create and run massplanner on at full power and without any lag, and then to buy only that many servers. For instance, if your server can support at most 4 massplanner (1 real 3 in a virtual machine) without any trouble than buy 3 VPS's, put 4 on the next one and put the last 2 on the third one.

The key is figuring out that magic number of MP instances on 1 VPS and then buying only as many as you need to get to your desired amount of instances.

Be warned: make sure you test them with all instances of MP at full power, because if you don't stress the system to its limit you won't have a good idea what it would look like under load. If you under anticipate the load, you'll end up with crashes and bugs of every kind. MP is not nice to RAM, even if your paging file is enabled, so make RAM allocation top priority.
 
  • Like
Reactions: Toz
Whelp, someone's rolling in the dough :p

The best way to have 10 concurrent instances of massplanner open at one time is to figure out the maximum number of virtual machines you can create and run massplanner on at full power and without any lag, and then to buy only that many servers. For instance, if your server can support at most 4 massplanner (1 real 3 in a virtual machine) without any trouble than buy 3 VPS's, put 4 on the next one and put the last 2 on the third one.

The key is figuring out that magic number of MP instances on 1 VPS and then buying only as many as you need to get to your desired amount of instances.

Be warned: make sure you test them with all instances of MP at full power, because if you don't stress the system to its limit you won't have a good idea what it would look like under load. If you under anticipate the load, you'll end up with crashes and bugs of every kind. MP is not nice to RAM, even if your paging file is enabled, so make RAM allocation top priority.

Lol thanks for the info!

Can you explain the paging file to me?

Also, what's to stop me from buying a dedi with 64 GB RAM and just allocating that acrossed 10 VMs?

Wouldn't that be more cost effective than paying for 10 VPSs monthly?
 
Lol thanks for the info!

Can you explain the paging file to me?

Also, what's to stop me from buying a dedi with 64 GB RAM and just allocating that acrossed 10 VMs?

Wouldn't that be more cost effective than paying for 10 VPSs monthly?
EDIT: That was longer than I wanted it to be.

The paging file is just a feature in windows that artificially extends RAM when you've used all of it by using your available disk space and storing the extra RAM information there. Because it uses the hard disk, it's slower than than it would be had all the RAM not been used up, and MassPlanner likes to crash once your RAM usage is so high that you starting using the paging file. Just make sure you have more than enough RAM.

You can buy a dedicated server and do exactly what you've said, but you need to do the numbers yourself to make sure it is the cheapest solution. I think the problem will arise in process cores. Whne running dedicated servers, the bottleneck resource is most often processor cores. Because you cannot share a processor core with a VM, you need to assign at least 1 physical core to each VM, but that would put a very high strain on each machine unless you're buying a dedi with 2 processors because as far as I know, the maximum amount of cores available in commercial dedicated servers is only somewhere around 15.

Off the bat, you'd need at least 3 cores for your host VM, so now we're down to 12. This means that you'll only be able to assign 1 processor core to most VM's, which is not enough for a host OS and a MassPlanner running at 100% load.

Your only option to run without hiccups is to either split the load between 2 VPS's, you don't need to pay for 10, you just get 2 and put 4 VM's in each of them. That or get a dedi with 2 processors. That'll give you at least 28 cores so you'd be good to go with 2 processors, but those servers are so expensive.
 
Last edited:
EDIT: That was longer than I wanted it to be.

The paging file is just a feature in windows that artificially extends RAM when you've used all of it by using your available disk space and storing the extra RAM information there. Because it uses the hard disk, it's slower than than it would be had all the RAM not been used up, and MassPlanner likes to crash once your RAM usage is so high that you starting using the paging file. Just make sure you have more than enough RAM.

You can buy a dedicated server and do exactly what you've said, but you need to do the numbers yourself to make sure it is the cheapest solution. I think the problem will arise in process cores. Whne running dedicated servers, the bottleneck resource is most often processor cores. Because you cannot share a processor core with a VM, you need to assign at least 1 physical core to each VM, but that would put a very high strain on each machine unless you're buying a dedi with 2 processors because as far as I know, the maximum amount of cores available in commercial dedicated servers is only somewhere around 15.

Off the bat, you'd need at least 3 cores for your host VM, so now we're down to 12. This means that you'll only be able to assign 1 processor core to most VM's, which is not enough for a host OS and a MassPlanner running at 100% load.

Your only option to run without hiccups is to either split the load between 2 VPS's, you don't need to pay for 10, you just get 2 and put 4 VM's in each of them. That or get a dedi with 2 processors. That'll give you at least 28 cores so you'd be good to go with 2 processors, but those servers are so expensive.

Most useful thing I've ever read on BHW.. add me on Skype pimp live:1244bec37ab36a57
 
You can run a dedi and then set up your own vms using hyperv. But if you run 10 copies of mp won't the mp licensing be expensive?

Why can't you run one copy on a dedi and have loads of accounts set up on it?
 
Some of the technical advice above is well past my pay grade and if I were to do what you want to do I'd be asking one of my tech partners/friends BUT I would not run them on one Dedi or anything else that should it go down all your MPers stop working.

Maybe in that tech jargon that was covered above but that would be the first thing I'd want to protect against.
 
You can run a dedi and then set up your own vms using hyperv. But if you run 10 copies of mp won't the mp licensing be expensive?

Why can't you run one copy on a dedi and have loads of accounts set up on it?
I think it's because MassPlanner doesn't support multi-threading, but I could be wrong. Either way, I think there's some sort of programmatic limitation that makes it unstable after (ballpark) 300-400 accounts.
 
Back
Top