Architecting, installing and maintaining your SAS environment

SAS Metadata Restore to alternate machine?

Senior User
Posts: 1

SAS Metadata Restore to alternate machine?

Wondering if anyone has tried this and has some insights as to if it is even possible.


This is in regards to 9.4 M2.


I am looking at disaster recovery type of situtations at the moment with our SAS cluster, particularly with how quickly we can get back up and running in a total loss situation of the cluster setup.


Question is, can a metadata backup be restored to another machine outside of the original cluster - so for example if we had a standby cluster is it just a matter of moving the files to a location that the server can get to them through the "Recover from Alternate Location"


The other element of that question would be if you haven't done a package export (lets say an OLAP cube and its measures/dimensions) at an individual item level, can you retrieve those from metadata backups without restore process taking place if (metadata backup) it cannot be recovered to another server  

Posts: 299

Re: SAS Metadata Restore to alternate machine?

The only time you could restore a backup from one machine on another, is if both machines are absolutely identical, "clones".

Paths, ports, machine names ... everything has to be the same. This approach is not recommended!


A good option for failover is using the Metadata Server cluster:


I worked with customer who restored a backup taken on one machine, on another, and it didn't work because

all the info about machine, install, config etc. is stored in config files and metadata objects.


Hope that help?



Super User
Posts: 3,866

Re: SAS Metadata Restore to alternate machine?

There is one situation where this is common practice and that is where there is an identical Disaster Recovery copy of your SAS server environment and you want to keep your DR metadata in synch with Production. In this case the SAS server setup is identical so it is quite OK to copy metadata even just using file copies. We do this daily in our SAS environment.

Ask a Question
Discussion stats
  • 2 replies
  • 3 in conversation