Tutorial on how to set-up a clustered scenario with mupx?


#1

At the moment, I have deployed my app on a single AWS instance with mupx, but I am facing problems with this. Once in a while (i.e. every 2-3 weeks), the server stops responding completely. Although the AWS console reports it as being OK, the server does does even react on SSH logon attempts any more. On mupx github, I was hinted that this might be a memory leak, but this is not sure at the moment.

I am still investigating this, but in order to prevent service abruptions for the future, I also want to set-up a fail-over scenario with a second machine that would take over if the main server is not reachable. Is there any good tutorial on how to do that in combination with mupx? I thought about integrating meteor cluster, but I don’t have any experiences with it yet, and I am open for all existing solutions out there. I would also appreciate any hints about things I have to take into account in my app once I implement a stand-by scenario. For instance, I heard that not all community packages are supporting this, and maybe I will have to adapt my app as well.


#2

you’ve heard of Galaxy, right? And it’s really not that expensive for what you’re getting. If it seems expensive let me ask you - what is your time worth?


#3

Yes, I’ve heard about it. But

  • no database hosting included (still need compose.io)
  • only US east coast supported (I need EU west)
  • doesn’t seem to be clustered by default either
  • no test-drive available
    For this, the price is too high for me at the moment. I will consider switching to Galaxy once the app really makes money.

#4

It’s really really easy to scale Galaxy. All your other points are still valid though.

I haven’t heard of anything outside of Meteor Cluster.


#5

Hello, will Galaxy come to China in the near future? Thanks