Select to view content in your preferred language

Query regarding planning Enterprise maintenance (daily exportsite backups) to get minimum downtime on publishing and editing services.

117
1
a month ago
AyushAggarwal
Emerging Contributor

We have a highly available, Mission Critical and Global environment that needs to be up 24*7, however we see that there are certain processes which are hampered during the ArcGIS Server exportsite backup. We did thorough testing on different environments and see that publishing a new service fails during an  ArcGIS Server exportsite backup fails each time.
 
After going through the below ESRI Blog: https://www.esri.com/arcgis-blog/products/arcgis-enterprise/administration/backup-basics-for-arcgis-...

The common thread that ties these together is time: backups should be taken at roughly the same time when new edits are not happening to ensure that data is being preserved in the process. This can involve down time, but at the very least should be accomplished when the system is not being used to edit, create, or delete information.

I realized that this issue is as designed. 

Our concern lies in the fact that we have a Highly Available, Mission Critical and Global 24*7 environment where we cannot afford to have a major downtime daily. (Taking into account that backup operations take more than 8 hours to complete as these environments are huge with bulk load of data), Also daily backups are an essential need in case of any Disaster Recovery.

Currently our situation is that we are unable to publish any new services for at least 8 hours daily which breaches our SLA and we don't know where to go from here. Any suggestion on what can be done to rectify or better plan the daily backup activity would be really great. We want to know if we are missing anything. 

0 Kudos
1 Reply
AngusHooper1
Frequent Contributor

If your requirement for resiliency is that significant then I would suggest making your design as simple as possible. I am guessing you do not have Portal in the mix as you are specifically talking about ArcGIS Server. You should have multiple distinct ArcGIS Server sites with independent storage tiers, authentication/access mechanisms, load balancers etc. You would then roll out a new service to each site independently of each other. Your backup and restore mechanism could then be within the storage and virtual machine tier. This is a long-winded way of saying that your uptime requirement is quite extreme and so your system design should be as simple as possible to support it using well-known fundamental processes such as storage backups or VM restores.

0 Kudos