Until recently we’ve had a backup process of bringing down sites daily for 20-30 minutes for incremental backups and Sundays for about an hour to do full back ups.
In the spring we implemented ER and surgical patient management applications that couldn’t tolerate the down time. We now do what I think is common, leave the sites running and either exempt some of the exec directories or ignore backup errors due to locked SMAT or DB files.
Our systems lead has now been warned by Healthvision not to back up the system without shutting down the daemons, processes, and threads on each site, citing the danger of DB corruption.
This has some in the head shed thinking we might want to look at a different interface engine solution that supports robust back ups with no interruption to service.
This all seems a little excessive to me…