Kbase P113026: Connection timeout by client on failed connect attempt to a database takes too long.
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
20/05/2010 |
|
Status: Verified
SYMPTOM(s):
Connection timeout by client on failed connect attempt to a database takes too long.
Client waits for approximately 51 seconds before a timeout occurs
Connect timeout takes longer in Progress 9 and OpenEdge 10 than in Progress 8.
Cannot reduce the connect timeout from the 4GL
FACT(s) (Environment):
Using the CONNECT statement from the Progress 4GL to connect to a database
Database is currently unavailable
All Supported Operating Systems
Progress 9.x
OpenEdge 10.0x
OpenEdge 10.1A
OpenEdge 10.1B
OpenEdge Category: Language (4GL/ABL)
CAUSE:
Bug# OE00134538
CAUSE:
This was caused by an intentional change in behavior in Progress 9. In Progress 8 we attempted to connect either 5 or 10 times, depending on the platform. In some customer scenarios this timeout was too short. In Progress 9 we attempt the connection 50 times.
FIX:
Upgrade to OpenEdge 10.1B01, 10.1C or later.
In 10.1B01 a new database connection parameter was introduced whereby it's possible to specify the number of connection attempts that will be made before a timeout occurs.
The new connection parameter is -ct <numRetries>, which needs to be specified for each client connection which requires a timeout other than that the default of 50 retries results in. For example if the default -ct 50 results in a timeout of 50 seconds before failing the connection (or causing a connection RETRY), then lowering this value will shorten the timeout accordingly.
Example where the connect retries setting is changed to 10 retries for the two databases:
$ prowin32 -db mydb1 -H myhost -S mydb1 -ct 10 -db mydb2 -H myhost -S mydb2 -ct 10
CONNECT -db mydb1 -H myhost -S mydb1 -ct 10 -db mydb2 -H myhost -S mydb2 -ct 10