I'm about to implement some sort of backup strategy myself this is very welcome advice. I do intend to use hard disk space, possibly to a network-mounted filesystem overnight.
I hope I can be forgiven for appending a supplementary question to someone else's thread, but: what would the command line incantation be to perform an incremental backup, assuming that's possible?
It is possible to do incrementals with probkup, however I wouldn't start with the software's capabilities. I think the first step should be devising a strategy that meets your business needs. Then look at the technology and its capabilities to determine how best to (or whether you can) implement your strategy. For example, two important metrics in business continuity are (a) how much data can you afford to lose (i.e. time elapsed/data changed the time of from your last recoverable change to the time of failure, aka "Recovery Point Objective"), and (b) how long you can afford to wait from the time of failure until time of recovery, aka "Recovery Time Objective". Obviously you want to minimize both but nothing comes for free. There was a good session on this topic at Progress Revolution 2011, with Brian Bowman, Adam Backman, Rob Marshall, and Mike McMahon:
http://download.psdn.com/media/revolution_2011/oe11/OE1102.wmv.
Like I said, you can do incremental backups in addition to full backups, but that approach may not be best for your needs, especially if your application is performance-sensitive. I had a client whose OLTP performance was mysteriously tanking every afternoon at 14:00. It turned out they were running an automated online backup every afternoon. Incrementals could have a similar impact, especially if you did them often. I suggest you look into OpenEdge Replication/Replication Plus (which is an extra-cost product) and after-imaging (which is a built-in database feature and should be enabled in any production environment). You can read about AI in the Database Admin manual that Tom pointed out, as well as in a presentation he created on the subject: (
PDF) (
video). AI can give you the recovery capability you want from incrementals, without the impact of frequent backups. Also, read about the After-Image File Management Daemon (AIMD), which helps AI extent management.
Is there a way to direct the backup output to stdout rather than a file? If there is, then you can do the gzip thing inline without having to compress in a second step.
I've never tried that, but it's not much work to add an extra line to your backup script to gzip the backup file. Typically you do this overnight so it shouldn't matter much whether it's one step or two (in a script).
RHD said it too.
