JDS Failover to Cloud Distribution Point

[bthomas]

I just setup a AWS Cloud Distribution Point with the thought that I would use it as a failover for our JDS in the event that we have a network or power outage (JSS is hosted externally) however it appears that there is no option to do this.

The only way I see to be able to have the cloud distribution point used as a non-primary/backup solution is to manually set the Cloud Distribution Point as the primary distribution point in the event of a failure of the JDS. This method isn't great either though because I don't want to store OS images on my AWS bucket and as a result if the JDS came back online and sync'd to the AWS bucket, I imagine it would blow out the images I have on the JDS.

Is the above thought process valid or am I missing something?

I looked into network segments but came to the issue that we are not a multi-site company but rather a company with one home office and ~100 field users spread across the United States. The only way I see having the AWS bucket in the mix with network segments is to make it the primary distribution point and to then set my JDS to be the primary for the home office network segment. I don't feel this is a great option though as it would significantly and unnecessarily increase our AWS costs due to all of the field users hitting AWS as the primary rather than the JDS.

TL:DR: What's the best way to setup a Cloud Distribution Point as a failover for a JDS in a single JDS environment with geographically dispersed users.

[rderewianko]

I don't believe you can use a CDP or JDS in failovers, there is a feature request on this https://jamfnation.jamfsoftware.com/featureRequest.html?id=1694 but as it sits its just in review
AFP/SMB/HTTPS (webdav) can be failed over.

[bthomas]

@rderewianko Thanks for the feedback. I have up-voted that feature request. It is incredibly frustrating that this feature is not built in as it makes having redundancy for a geographically distributed workforce (work from home) next to impossible.

[rderewianko]

I agree, its why I only run webdav and SMB distros.

[bthomas]

But correct me if I am wrong, that would not provide functional failover for users connecting via the WAN. They would still only be hitting the WebDav server they are assigned to. The failover, as far as I understand it, only applies to the SMB/AFP distros and in my environment most of the users cant connect over AFP/SMB unless I was crazy enough to put a globally accessible SMB sharepoint out on the internet.

Another possibility, however complex, would be to host a JDS in AWS and use DNS failover to flip over to the AWS instance of your JDS if the DNS service detects an outage at the main JDS. The JDS inside of AWS could have its own unique FQDN and name but the DNS would just treat it the same as the primary JDS in the event of failover. I am going to setup a JDS in AWS and give this a test with out failover DNS feature. I'll report back with any failures or successes.

[rderewianko]

There's a neat trick where you can run a FileShare DP just in Webdav mode.

@brysontyrrell references box.com as an external webdav dp but you can use any other provider.
https://bryson3gps.wordpress.com/2014/09/12/using-box-as-a-distribution-point-the-9x-version/

Internally we use AFP/SMB externally we have a webdav point

Our flow is like this: Hit internally, it fails -> go to external (if casper hasn't figured out you're outside the office) (or if you're on a guest network)