Kbase P100545: Fathom Mangement: DB_AgentCrash with error 9742
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
2/17/2009 |
|
Status: Verified
SYMPTOM(s):
Alert Name: DB_AgentCrash
The admsrv.log records the Agent failure (9742) (9489) (9615) (10472)
AgentCrash: Database Agent, <agent name>, crashed! (9742)
Fathom is unable to retrieve trend data from database: <Database name> Ensure the database is running. (9489)
Invoking E-mail action. Action: <name of the e-mail action>, Resource: <name of resource that caused the alert which invoked this action>, From: <e-mail sender>, To: <e-mail recipient> (9615)
DBAgentReadError: A problem was encountered getting data from database <databasename>. The data will be ignored. (10472)
A short time later, message 9490 is registered confirming reconnection
Fathom is now able to poll the database: <Database name> (9490)
Around the same timestamp the database log file shows the agent loosing connection to the AdminServer
FMAGNT: Connection to Admin Server lost. (8848)
The agent immediately recovers the connection to the AdminServer
FMAGNT: Registered with Admin Server. (8846)
At the same timestamp there are error 3695 in the database log file where client sessions are logged out
Signal 1 received: not expected. (3695)
FACT(s) (Environment):
Fathom Mangement
All Supported Operating Systems
OpenEdge Management
CAUSE:
From the evidence (presented above) it appears that there was a Network "glitch" around this time which was recovered.
Other users, received error :
Signal 1 received: not expected. (3695)
Signal 1 is restart, this was for the normal usr, not for the agent but does further support the assumption that there was a Network "glitch" around this time rather than a problem with Fathom Mangement.
They too appear to have been able to re-connect within a very short span of time.
FIX:
No actions are necessary unless it doesn't recover. While the dbagent for the database is not running, no polling of that database are taking place. The database itself keeps running, it is just not being monitored. When the agent is restarted, polling continues. The dbagent has re-connected and according to the current status, is running so therefore polling has resumed.