|
|
|
Home » U++ Library support » U++ SQL » How to manage the access to a Db on an unstable connection
| How to manage the access to a Db on an unstable connection [message #50398] |
Tue, 16 October 2018 11:51  |
Giorgio
Messages: 218
Registered: August 2015
|
Experienced Member |
|
|
Hi there,
I have developed an application used in the shop floor of a manufacturing company. Some clients collect production data and save them on a PostGresql DB. Initially, clients were wired to the LAN and there were no performance issues. The connection to the DB is managed in GUI_APP_MAIN: the connection is established and then I use the global SqlSession variable SQL throughout the whole application. Recently, some clients were added and connected through a wireless connection, as it was difficult to have cables running all over the shop floor. The wireless connection is unstable by definition, so it happens frequently that the connection is lost. Basically, there are two scenarios, when the connection is lost:
1. The user is just on a "view" status (a SqlTable filled with data) without issuing any sql statement.
2. The user is performing an action that requires an insert/update/select statement to be issued.
In the first scenario, if the client can restore the connection in a reasonable time, the user does not notice anything; otherwise everything on the SqlTable disappears: the user can cleanly exit the application and after restarting it usually everything works.
In the second scenario the application seems "freezed" (it is waiting for the DB server to respond, so it "hangs"). Sometime the "freeze" lasts only few seconds so the user just waits. In other cases the "freeze" lasts for more time (either because the connection with the db is completely lost, or because it takes sometime to the client to get back online), so the user kill the application.
For the above reasons, I have to rethink the way my application connects to the database, but I am a bit confused because is the first time I am in such a situation.
Basically I think the flow should be as follows:
-the application starts and connect to the db;
-after the initial select statements (to fill in all the necessary data) the application disconnects from the db;
-any time an operation that requires a db connection is performed (e.g. adding new data, when the SqlTable runs the Query() method, etc.) the application connects and - after the operation is performed - it disconnects;
-during each connection, an appropriate timeout is set and if the connection is failed the user is informed (so the user does not thing that he has to kill the application);
-if possible, the global SQL variable is used in the connection/disconnection actions.
So that is my idea, but I would be really happy if anyone want to share is ideas and experiences in developing a DB application in an environment where the connection is unstable.
Regards,
gio
|
|
|
|
|
|
| Re: How to manage the access to a Db on an unstable connection [message #50412 is a reply to message #50400] |
Tue, 23 October 2018 10:46  |
Giorgio
Messages: 218
Registered: August 2015
|
Experienced Member |
|
|
Hi Zbych,
thank you for your suggestions.
My routines that deal with the database have all the same structure:
int SaveData::TotalDrums(String c)
{
int retval;
SQL.ClearError();
try {
retval = SQL % Select(SqlSum(DRUMS))
.From(MYTABLE)
.Where( MYFIELD == AsString(c) );
} catch(SqlExc) {
retval = -1;
ErrorOK(t_("Error: ") + SQL.GetLastError());
}
return retval;
}
Do you think I should catch also something else?
Is there on the upp site some documentation on ConnectionOK etcetera? I found almost nothing.
Regards,
gio
|
|
|
|
Goto Forum:
Current Time: Sat Apr 19 03:08:00 CEST 2025
Total time taken to generate the page: 0.02325 seconds
|
|
|
Members
Pages
Help
Register
Login
Home
