We've just upgraded from Server 2008 R2 / Backup Exec 15 to Server 2016 / Backup Exec 20.3. We have 2 sites and duplication jobs to send backups between sites for DR.
What we did on the old setup was store the catalogs on a Windows DFS-R folder (that was local to each Backup Exec server), then point Backup Exec to this folder for its catalogs. With duplication, that meant that any backup job from either server was immediately restorable from itself and the other. In our primary site CASO and Backup we co-located on one server.
Now, we have separated CASO and Backup in the primary site. During install of Backup I chose to connect to CAS and store catalogs on CAS, with unrestricted catalog access. I've yet to upgrade the other site so can't test what happens.
What I want to be able to do is backup locally and duplicate to the other server, then if one site fails I can just immediately restore from the surviving site. Some of our services (email, database) only exist in the primary site, hence the need for duplication.
If the site with CAS fails, what happens to the backup server in the other site, if catalogs are stored centrally?
If used centrally managed catalogs and you lost you CAS, you need to do inventory and catalog operation on the survived MMS server for its local storages. Or you need to switch from centrally managed catalogs to replicated or distributed catalogs https://www.veritas.com/support/en_US/doc/59226269-99535599-0/v53927828-99535599
Thanks. Just to confirm:
Sites are connected via 100Mbps fiber. Don't know if Veritas deem that "high-bandwidth".
If MMS 1 is set to centralised, I cannot perform restore opteration if CAS is down (for maintenance for example)
If MMS 2 is set to centralised, I cannot perform restore operations if CAS is unavailable (maintenance or WAN or Site A failure for example)
If MMS 1 and 2 are set to distributed or replicated, I can perform immediate restore operations for all data on any MMS
MMS 1 with centralised catalogs backs up email data and duplicates it to MMS 2 - this data is cataloged on CAS only
MMS 1 with distributed catalogs backs up email data and duplicates it to MMS 2 - this data is cataloged on MMS 1 only
MMS 1 with replicated catalogs backs up email data and duplicates it to MMS 2 - this data is cataloged on CAS and MMS 1 only
MMS 2 with centralised catalogs backups up other local data and duplicates it to MMS 1 - this data is cataloged on CAS only
MMS 2 with distributed catalogs backups up other local data and duplicates it to MMS 1 - this data is cataloged on MMS 2 only
MMS 2 with replicated catalogs backups up other local data and duplicates it to MMS 1 - this data is cataloged on CAS and MMS 2 only
So for example, Site A totally falls over - I need to just be able to immediately restore email data in Site B as if it was actually backed up on MMS 2. Without having to catalog the duplication jobs data on MMS 2 first.