Method Community

I am having an issue inserting data into one of my tables via the API.  The CoursePayments table is telling me when I try to insert data that the "Table does not support add."  When the call happens it is working about half the time, which is concerning since the data looks almost identical between calls to insert the data.  I have checked the forum and noticed there was a bug lastyear involving not being able to add to tables.  Is there anything you can do to check if this issue is coming up again.  

I don't believe the issue has come back.  I'm assuming CoursePayments is a custom table you created, is that right?  I created a table with the same name, added some fields and everything seems to be working as expected.  I'm guessing there may be an issue with your specific implementation.  Can you share any additional details about your table, such as fields and which are required?    

We were adding and deleting a field to the table that "broke" the table earlier in the day. We broke the table by trying to link a filed through too many times I think. We tried to link a field from courses Table->Enrollments Table->CoursePayment Table->CoursePaymentsTable. But this bug occurred after we had finished trying to add that field and had changed it back to its previous state where everything was working.  I am curious if you use a load balancer to handle incoming calls from the api.  It felt like the call went through sometimes and not others randomly, so I was curious if one server might have had a version of that table from the morning when it had the broken field, and so that is why the call failed some of the time. 

That's an interesting idea.  The API itself does run behind a load-balancer and is deployed on multiple servers.  However, the servers themselves should have the same code deployed (and there are controls in place to ensure that is the case).  The individual servers can be up or down at various times throughout any given day for a variety of reasons, but the load-balancer will account for that and stear traffic away from servers that are down.  Each of the API servers access the same database (which is not load-balanced but has other mechanisms in place to ensure durability).  Theoroetically, the behavior you saw should not be possible.  However, in practive, maybe there was something wrong with one of the API servers and the load-balancer did not know about it and continued routing traffic to it.

Please keep us posted if you are still seeing this behavior, or if you see it again in the future.  I have confirmed that all servers are online and functioning correctly at the time of this post.