I have successfully created a backup of my production Environment using the WebGISDR tool. I have been told I can restore to a new environment in the same network aliasing the new servers with host files. Has anyone successfully restored this type of configuration?
Are you running into an issue with that workflow? The main thing is that you need to make sure that the internal URLs that your environment is configured with don't resolve to the original primary environment. As a basic example:
The right hand side is production, the left hand side is a new standby environment. The web adaptor in the second environment has it's own hostname, but you're telling m2.domain.com to communicate with the web adaptor machine through an entry in the etc\hosts file. ArcGIS Enterprise is configured to use the "enterprise.domain.com" FQDN, but those requests will be routed to the standby site based on the etc\hosts entries.
Another approach, if you can manage the downtime, is while configuring the standby, configure your reverse proxy to send traffic to the standby machines, just until you configure the environment. Once that's done, you can redirect traffic back to your original environment.
Another key point is that the internal URLs, (privatePortalURL and the admin URL used in federation), must resolve to the standby environment via either etc\hosts entries, by configuring them to point to the machines directly, (m2.domain.com:6443/m2.domain.com:7443), or by having separate components that can handle that traffic.
Thanks for the quick reply Jonathan.
I have my stand-by environment set up and using etc\hosts to prevent any of the components (Portal, Server, DataStore) from resolving to the primary network. I do have a support ticket for this issue and hope to have a solution at some point. I was just reaching out to see if any other users have had success with this workflow.
Hi, E.J.
I realize some time has passed since your post, but am curious if you were ultimately successful in your efforts with the webgisdr utility?
We are attempting to follow the same process. We were successful using a test that consisted of three servers (primary single enterprise server, secondary single enterprise server, dedicated server for data). Since we were successful with our testing, we decided to re-configure our enterprise environment and upgrade to 10.7.1 at the same time. We have gotten our primary enterprise server set up, we have a dedicated data server in place, and have gotten a secondary server set up using the etc/hosts file.
Where we have not been successful is in being able to perform a consistent restore from the primary to the secondary server. We would like to have a reliable, dependable process. We would also like to be able to automate the process.
Would love to hear if you have been able to complete, repeat and automate your backup/restore process.
Hope you are staying safe and sane. Best, Dixie.
Hi Dixie,
We have been successful in getting the WebGISDR to backup and restore to a stand by environment and even have it running as a scheduled task. The key I believe is when we did a manual backup of the data store from the Production environment to the Stand By Environment.
I believe the documentation for that is here<https://enterprise.arcgis.com/en/portal/10.7/administer/windows/data-store-backups.htm>. Once that occurred, the WebGISDR process began to work. I think it synchs the datastore so the rest of the backup will flow.
E.J. McNaughton
Oh, wow! Thanks for the information, E.J.!
So, if I understand you correctly, you ran the manual restore of the data store to your standby environment and then you ran an import of the full webgisdr backup?
Thanks again for the help! Best, Dixie.
Thanks again for the guidance, E.J.!
Now I seem to have a bootstrapping problem with restoring the backup. I cannot get the secondary data store to recognize the primary data store's backup.
Did you use a share as the location for the primary data store backup and then use that same location for the restore? Did you register that location with the secondary data store prior to issuing a restoredatastore command?
Thanks again for your help and patience (^-*)!
Best, Dixie.
Well, dang! I will have to continue to pursue help via ESRI support. We have a pending support request about our issues with trying to get a successful restore on our standby server.
Perhaps our bigger issue is centered on the data store because I could not get the datastore restore to work from either a local directory or a share.
The errors were either 'Failed to create relational data store database '{0}'.' or 'Could not find a valid backup to restore the data store.'
Thanks again for all your patience and advice.
Best, Dixie.