I have a project running where server address URL is behind a domain name. The server is going into overload and people are having difficult in sending data back to it (they have already downloaded the blank form and filled it). One possible solution I am thinking is this: I setup a new empty aggregate server, load the same form onto it as current server, and switch the server IP behind the domain name. This means that the new IP/Domain name will be updated in DNS in few hours, and people will start hitting the new server aggregate instead of the old server.
My concern is this: Once the current server is offline and all traffic is rerouted to new server, how will I access the aggregate via web interface (and/or subsequently take backup via briefcase)? Because aggregate is configured to work over domain name URL, so old server aggregate will reroute me to URL (which now will take me to new server). Same is the case with briefcase: When i try the pull, how will I access the old server? IP address would not work, right?
Is there any way to modify the aggregate base URL to access data via IP address or something?
@LN, this is in continuation of the other thread you replied to (Massive dataset - Need optimization advice). Your advice there was helpful, many thanks. I will send its reply soon as well. However, please guide me here about the above, if possible. Thanks!
Hi @Saad, you're correct that you'll need to reconfigure Aggregate to change how it is accessed. Here's a previous thread you were involved in that describes how to do this: Changing Aggregate IP to URL! You can do the opposite of what you were describing there and enter the public-facing IP instead of the domain name.
Ahhh yes, I forgot about this one. Perhaps I am too sacred to lose the data that my thinking has been defensive and not following the same old good-hacking pattern ... Let me try that.
Just to be sure: I will only be changing the front end, and the aggregate database will remain intact, right?
I understand. Dealing with server instability is really stressful. I hope you're close to a resolution.
Yes, that's exactly right. As always, do make sure you have a complete backup but if you review @yanokwa's post at Changing Aggregate IP to URL you should be able to make the change without risk to your data.
That's the main issue here actually. The server is so choked and so heavy that I cannot even take a backup!!! That's why taking the really really no-risk steps.
Oh no! Is Tomcat the bottleneck? Can you make an AWS-level backup?
Now that when I try to think innovative, I can also make a clone of server on AWS and pay around with the cloned server... yet another option.
Yes. Tomcat is really giving up on me. It's not really an ODK/aggregate issue, but tomcat 8 is unable to handle the load now, even though I have extra RAM and CPU. I tried to find some alternate webserver, but so far I have not found any. Nginx doesn't support Java, else I would have tried that.