I am using 12.8 with Workgroup database, Windows.
In Progress Explorer shows that After Image Writer: N/A - while I can use rfutil to see that the ai is running. Is this normal?
The After Image Writer is not the same as having AI running. You will want to start an AI Writer process. You will want to add this to your startup scripts along with the BI Writer process you should already have in place (probiw)
Progress Documentation These two daemon processes take load away from the DB for writing AI and BI notes.
The writer processes (AIW, BIW, APW) are performance enhancers, as
@jdpjamesp said. However they are exclusive to the Enterprise RDBMS license so you cannot use them. The only helper process available to you is the watchdog (prowdog command), which you should use.
My archive folder has too many files past-dated.
"Too many" and "past-dated" are value judgements. I have had clients use months-old AI files. Think about why you might use archived AI extents. It isn't always about disaster recovery, and needing only the files since the backup you are restoring. They can be useful in other scenarios, like forensic investigations, e.g.:
- Your boss wants to know the state of a record on the 15th of a month, several months ago, but all you have from that far back are two monthly backups. Without AI files, you can't answer the question. With AI files, you can restore the older backup and then roll forward all the files up to the 15th of that month. You can roll forward an entire AI file, or up to a point in time, or even up to a single transaction.
- Your boss wants to know when a particular record last changed, and who changed it, and you aren't using auditing. With archived database backups and AI extents, you can answer that question. Without them, you can't.
In a workgroup database, which I would expect to be small and slow-growing, backups and AI files should be quite small. AI files compress very well, making their footprint even smaller. And since you are on 12.8, you can take advantage of the new (since 12.5) backup compression using the -com option of probkup. You can also experiment with -comlevel 3 or 4 to see how the compression ratio and backup duration compare.
For backups, I do "probkup online <source><destination>" - is this best practice or is there something better with regards to AI?
Backups are important. It is just as important to monitor them on a daily basis. Did the backup run? Did it succeed? Is the backup file size what you expect? Was it being safely copied to an off-site DR server, along with your archived AI files? Ensure that your -aiarcdir is in a file system local to the source database, not a network file share on a remote server. You need a scripted process to copy the files from there to your DR server.
Some organizations back up their databases religiously but know little about restore. Or they find out too late that the backup files they thought they had are no good. Ensure that you incorporate regular database restores, and you have scripted routines to roll forward archived AI files against the restored databases, so you aren't trying to figure out how to do it in the heat of a disaster.
