Currently, we are log shipping and restoring to a backup server as our primary DR. We are also taking full, diff, and tlog backups as a secondary DR precaution. We are running the backups late at night but our operation is 24/7 and the backups are causing disk latency while running which is slowing the application. I was hoping I could switch to take the backups from the backup server (where the log shipping is happening) but apparently you can't backup from a database in standby mode. Other than adding more disk power, which I'd prefer not to do because they are fine for the normal operation of the database 22 hours of the day, what other options do I have?
These databases are Standard Edition. In the medium term, we are looking to move to AWS RDS from EC2 instances and make use of the replica feature but I'm looking for some shorter term options.
Thanks!
This is most likely an IO configuration issue - and you should consider increasing the disk performance to resolve this issue.
Is the storage on a SAN? If so - what kind of SAN and how is it connected and configured? Do you have separate drives for data and backups or is all data on the same volume?
Hi Jeff,
The drives are EBS volumes in AWS. For the database in question, there is a data drive and a separate drive for the tlogs. The backups read from the data drive and write to a different drive. The latency is read latency and happens during the backups. The drive is a 2TB EBS volume which gives maximum IOPS for non-PIOP drives. We could configure a RAID or simply add volumes to the database (put some large tables or indices on a different drive) but since the performance is fine 22 out of 24 hours per day I was hoping there was another route to take.
Thanks!