Hello all, Just a quick update, at the NEPUG meeting this past Feb.
2008, I gave a presentation on AI, included was a how to on how
aiarchive works. Bob Brennan will be posting that PowerPoint soon.
Fascinating how the Progress folks wrote this feature GREAT JOB!!
Message was edited by: Bob Bolduc
PowerPoint explains it very well. Would like to see something put together by Progress as an online presentation, if possible, or atleast discussed at Exchange.
Yes, I use it and it is A Good Thing.
Particularly if you do not already have good scripts of your own to handle the after imaging process. (I do have such scripts -- but they're UNIX oriented and with smaller scale Windows users I'm finding the built-in features more suitable.)
I do, however, have a complaint. The naming convention stinks. I'd like control over how the files get named. I generally only want the base name and the sequence number. Plus maybe a user supplied reference of some sort. The date and time are, IMHO, just noise. Being able to specify a "mask" with placeholders for various variables like printf() or the formatting for the UNIX "date" command would do the trick nicely.
One other point... "official" PSC presentations are nice. Especially to discuss the gory details of internals that us mere users cannot know about. But PSC doesn't use this stuff in the real world -- if they did they would have a more reasonable file naming convetion ;). I'm personally more interested in the experiences of actual customers with real production systems when it comes to a feature like this. FWIW there is an Exchange presentation on the topic -- OPS-11 (Full disclosure note... Adam & I work together.)
Ok, one last point. The roll-forward process needs some work too. We need to be able to roll forward multiple extents without starting & stopping and going through the redo phase with every extent. And it would be nice if it just skipped over invalid extents. IOW there should be a command along the lines of:
rfutil dbname -C roll forward file-list
or maybe even:
rfutil dbname -C roll forward dir-name
If some of the extents in the file list (or directory) are inappropriate (maybe they're from the wrong db or they were archived prior to the backup that they're being rolled forward against then they should just be skipped (as they would be if done manually and one at a time).
we also use it. the filename convention stinks
I also concur that a roll forward from directory is a very useful tool to have.
Hey, Tom! I agree with you
Pretty soon they'll be calling it a movement...
Some kind of movement, anyway ...
Now if we could just get someone to set it to four part harmony we'd be getting somewhere
With only two singers, that would require Tuvan throat singing or the like ...
what are you talking about ? We have already formed the PAFFF. (the Popular Archive Filename Freedom Front)
I'll bet that I can come up with some pretense or other to get Tom Harris and Rich Banville to join in and that makes 4!
And with enough beer it won't matter how harmonious our singing is! But don't worry, I'll have a camera handy to make sure that the agreement to fix the filenames is properly attested to.
doesn't the singing get better with more beer ? Soon we could be topping the US charts.
Careful, if you keep that up you're going to get this thread shutdown for unsportsmanlike commentary...
a question: I am using -aiarchive, is there any way of forcing the ai manager to make a switch and copy the ai file ? (keeping the same filename layout)
Message was edited by:
Julian Lyndon-Smith
What happens when you manually run:
_rfutil dbname -C aimage new
well, I never. It worked.
Thanks!