Kbase P48660: Error 1068 generated by WebSpeed agents.
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
11/13/2008 |
|
Status: Verified
SYMPTOM(s):
Database clients are WebSpeed agents.
SYSTEM ERROR: rmmak called during resyncing (1068)
When error occurs, several WebSpeed clients crash with 1068.
Database does not go down after error message.
At the preceeding login of the WebSpeed agent concerned, the following message is written to the log file
Received RECONNECT from WTB
Stack trace from _progres reads:
rmCreate
dsmRecordCreate
procr
kmkadd
FACT(s) (Environment):
All Supported Operating Systems
Progress/OpenEdge Versions
CAUSE:
There were more than one database in the environment. After ONE of these databases terminated abnormally, it was re-started without first shutting down the WebSpeed Agents that were connected to it. So when this database was re-started, the WebSpeed agents were still connected to the PREVIOUS instance of this database startup and threw the "SYSTEM ERROR: rmmak called during resyncing (1068)" as they crashed. The database remained up and running and the WebSpeed Agents were able to re-connect.
"rmmak() returns the recid of the record it created or a negative value that indicates an error (lock table full for example). In this case, when the WebSpeed Agent is trying to write to the database and there was a problem, because it was started against the shared memory of the PREVIOUS database startup, so the current shared-memory of this datatbase cannot handle the data. We then try to resend it,
failed, and return the 1068 error informing that we were trying to resync the data but could not.
FIX:
Make sure that all Agents have properlly terminated before re-starting the database.