GitHub user marving123 added a comment to the discussion: Disaster recovery in cloudstack
@TheKunalSen I would also add to test the DR flow with one or two VMs before planning it for all 100 VMs... For the NAS Backup and Recovery option, instead of using one NFS repository over the WAN, you can keep a local backup repository at each site and sync the backup data between them. CloudStack 4.22 can then create the VM in the other zone from the backup. This should also give you a real idea of the recovery time. If restoring one large VM takes 30-40 minutes, then this may be fine for normal DR, but probably not enough for websites where you need failover in a few minutes. For those critical sites, I would sugegst keeping the VM/application already available at Site B and replicate the database/data separately. Then Cloudflare can fail over traffic without waiting for the VM restore. I think doing this with one test VM first will make it much easier to decide which VMs need backup-based DR and which ones need a warm standby. Marvin AccuWeb.Cloud GitHub link: https://github.com/apache/cloudstack/discussions/13146#discussioncomment-19456698 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
