I do auto backups, do not depend on customer to do the backup. The customer however does do backups to an external drive, daily. Makes them aware that backups are important. Onto the same server I have 5 folders named Mon, Tues, etc. Set up a scheduled task to run an R:Base program that determines from the system date which folder to copy the database into. It is done in the wee hours of the morning when no one is around.
Also set a timeout in your startup program to close the database if no action in a couple hours. Just in case someone forgot to logout. My 2 cents, Bernie Lis ----- Original Message ----- From: MDRD To: RBASE-L Mailing List Sent: Monday, April 26, 2010 5:45 PM Subject: [RBASE-L] - RE: Strange problem? Hi Tony Thanks, I will check into that... I really think the error was between their ears! Marc From: A.G. IJntema Sent: Monday, April 26, 2010 4:34 PM To: RBASE-L Mailing List Subject: [RBASE-L] - RE: Strange problem? Hi, I have followed your discussion about the loss of data a little bit. Instead looking for a technical solution I have another suggestion. In all my applications I have a steering table which only contains just 1 row and it keeps track of important data. For instance: Some reports are financial relevant and they contain a sequence number which is incremented by one for every new report that is created. Accountants love it, because they are able to check if all the reports are there. Another thing you can do is add the report number to the row which is reported. In this case you are able to reconstruct a report afterwards and it can be proven that financial data have not been changed afterwards. Accountants love this feature even more as you can imagine. Another possibility is to keep track of the backup date and expand the database name with a sequence number of the backup. In this way you achieve two goals. First the backup database is not usable without renaming it. Secondly you always able to keep track if and when a backup has been made. Hope this view helps a little to tackle your problem. Regards Tony IJntema The Netherlands From: [email protected] [mailto:[email protected]] On Behalf Of MDRD Sent: maandag 26 april 2010 15:06 To: RBASE-L Mailing List Subject: [RBASE-L] - Strange problem? Friday I had the 2nd user this year call and say that Rbase lost a weeks worth of data. I am sure RBase had nothing to do with this and I explained that it is impossible for a weeks worth of data to get deleted without corrupting the data base. The user said there were no errors on Autochk. I told them it appears that somehow they connected to an older backup copy of the data. They said they were missing Customers, transactions, recap stats everything from X date on. I explained how PK's, FK's .... and such I would have to delete the data in reverse order such as I can not delete a customer until I deleted their transactions because the data is linked... The user seemed to understand but is there any way I can prevent this or make it easier to prove it was a staff error or computer error not an RBase error? Another strange problem, a user called and they had transactions in the Daily table for customer 0 which does not exist in the customer table, there are PK's and FK's on those tables. I told them I thought they had to have some kind of short on their network because we can't even enter a transaction for customer 0... so for it to get in there there was some kind of computer hiccup. The normal start up routine is to run Autochk, if it passes make a copy of the DB files called Bck, then Zip them up with the day of the weeks name on it such as Monday.zip. Strange thing is they swear they do this everyday yet some of those files were dated March 17. I am now going to add some code to check the date of those files and if they are too old give them a warning. Any suggestions? Marc
