There is no way that a backup can destroy data so something else is going on.
When you create a record, is there a gap in the serial numbers that shows the number of missing records?
Any cascading deletes set up in the relationships?
I have no deletes in my relationships. I only allow creates.
I use the relationship database as a tracking tool as the order processes along. Each step is a new process. It has Name, Date, Completion Date and Notes. A new record was created yesterday for each order/item processed.
I do not assign serial number or record number in the database. Should I? (I guess I can see if I'm missing any records that way)
1 of 1 people found this helpful
There is only one scenario that fits the description of your problem (as far as I can see).
You should look for an 'Order' database on your own machine, which you inadvertently logged into yesterday.
Records do not disappear from served databases and particularly from their backups - so, you must have entered them into another (possibly test) system, which is probably located on your own machine.
We have all done this just once …
Search your machine with spotlight, using the name of your main database (Orders) and then look in any files found for the missing records - I'm certain that there are there.
Best wishes - Alan
1 of 1 people found this helpful
I saw something similar in a legacy database which was moved to 12. The data were actually being over-written. Where or how, we don't know, but someone did something in the system, and the data were gone. Troubleshooting was going to cost too much for the client. However, I like Alan's comments.
I leave my computer on 24/7 and have never closed the database yet. The relational records are still missing.
This isn't the first time this has happed to me (been happening last 2-3 weeks). Finally got feed up with it and went on Forum this morning.
Just got back from Cardiac Rehab at 12:30 P.M. my time. (Above 2 lines were entered around 9:00 this morning)
OK, now I'm more confused (but happy). The records are now back! I can see them. And my Noon bakcup has the records.
I have notice that it takes the server physically longer to write to the database. I have 2 computers I use. If I enter data on 1st comperter, it can take upto 10-15 secords before I can see it on the 2nd computer. But 6 hours later? It still doesn't explain me not seeing the data on the 1st computer that entered the data then lost it.
I'm using FM11. Haven't upgraded to 12 yet.
I have my RAM at maximum on the server. And flushing every 1 minute.
Could that my slowing my server response time?
You mention "relational data." Are you entering the data through a portal? If so, not committing the parent record will cause all of the portal data to be deleted. This is considered a feature.
Another occasional problem is users entering data when they are in find mode, instead of browse. It can look like they are creating records, but the are actually creating find requests. There is a warning about this after about 10 records.
IT sounds as though you are working on a very small network, possibly with peer-to-peer sharing enabled for one client-type machine to work as server to the other, which can be a data-loss-risky proposition.
Even if you are running FMServer, also keep in mind that data isn't entered on the server file until it is committed by the client entering it. I have seen people walk away from a huge data-entry session but leave their cursor inside a data field, only to be stunned when Server disconnected them hours later without the data being committed.
That obviously didn't happen to you this time or it couldn't have reappeared. But that might be one reason new data seems to take so long to appear on both machines.
Our system is global. So my server is actual located in another state than I'm in. The server severs up North America, Asia, Europe, South America, basically the globe.
This morning I have only lost some of my records (data). Most are there.
The last comment I received sounds like something I am doing to myself. Yes I am using a master orders database and entering records into the relational database for my process scheduling.
The master orders database has portals I entering into. In fact 33 portals. Each portal for each process. I find the order in the master database and enter the schedule on the portal databases (relational).
I import SQL data into the master database and schedule through the relational databases. I never really enter data into the master except the import from SQL. I not sure how to commit in the master database. Is that hitting the enter button when I'm done or is that entering some "kind" of data in the master after I do the relational databases?
Anything that commits the master record will commit all of the portal data.
Anything that prevents the master record from committing will delete all of the newly entered portal data. This could include deleting the record, reverting the record, or getting disconnected from the server before a record commit.