Резервное копирование и восстановление данных

Service and Data Backup and Recovery

RackCorp UMS Application

All RackCorp UMS servers are deployable and rebuildable from the API / Portal Controller which makes for fast re-deployment.  Base RockyLinux OS needs to be deployed, and a script run to register the server on the RackCorp UMS Administration System.  once run, a server will be available for installation:

Once a server finishes installing, it will search for other same-class servers within the same node.

No nodes found (new install or complete loss of data)
Service will be operational, with no data (i.e. no cloud data for users, no email for users etc).  Option exists to manually copy ZFS datastore from backup storage to the appropriate location.  Other nodes will then pick up on this and commence synchronisation.

Another node found (IMAP role)
Download of data will automatically begin to synchronise the IMAP stores

Another node found (CloudDB role)
Download of data will automatically begin to synchronise the CloudFIle DB stores (meta data).

Load Balancers, Firewalls, UMS API, UMS Portal applications all generate daily dumps which are copied to any available S3 / CIFS/NFS storage.  These services contain data that typically does not change frequently.

Bootstrap:

# After base OS is installed, execute the following as root:
cd /tmp
wget https://api.XXXXXXXXXXXXXXXXXX.com/install/bootstrapcore.sh
sh bootstrapcore.sh

# This may take about 30 seconds to register in the portal.  Node personality can
# then be changed to its purpose (i.e. SMTP, IMAP, LoadBalancer, etc)
# Portal can then be used to deploy application to this server