OBIEE agent failed

961614

Hi,

Obiee agents get trigger based on VB script. The files stored in FTP server. But today one agent was failed with below error.

I got below error while running manually.

-----------------------------------------------------------------

Global Error: [nQSError: 77006] Oracle BI Presentation Server Error: Authentication failed.

Error Codes: WKY9UVRD:

Authentication Failure.

Error Codes: IHVF6OM7:OPR4ONWY:U9IM8TAC

Location: saw.connectionPool.getConnection, saw.rpc.server.responder, saw.rpc.server, saw.rpc.server.handleConnection, saw.rpc.server.dispatch, saw.threadpool.socketrpcserver, saw.threads

Odbc driver returned an error (SQLDriverConnectW).

State: 08004.  Code: 10018.  [NQODBC] [SQL_STATE: 08004] [nQSError: 10018] Access for the requested connection is refused.

[nQSError: 43113] Message returned from OBIS.

[nQSError: 43126] Authentication failed: invalid user/password. (08004)

The below one is Scheduler log

-----=-==================

[2017-01-11T05:00:01.000+05:30] [OracleBISchedulerComponent] [ERROR:1] [] [] [ecid: 00id5k_UUcQ6mJ45JVXBif0000lO00007F,0:1:3] [tid: 3728]  [nQSError: 77006] Oracle BI Presentation Server Error: A fatal error occurred while processing the request. The server responded with: Inconsistent job ids.  Scheduler job id 105 points to ibot  [[

/shared/SCHEDULED REPORT/Netbanking/15G_15H_Report" with job id 444.  A catalog update has been

made without updating the scheduler tables.  The Administrator should consider deleting the old scheduler job.

Error Codes: C8TDLG6P

Location: saw.delivers.rpc.getMetadata, saw.rpc.server.responder, saw.rpc.server, saw.rpc.server.handleConnection, saw.rpc.server.dispatch, saw.threadpool.socketrpcserver, saw.threads

.

Error Codes: AGEGTYVF

]]

[2017-01-11T05:00:01.000+05:30] [OracleBISchedulerComponent] [ERROR:1] [] [] [ecid: 00id5k_UUcQ6mJ45JVXBif0000lO00007F,0:1:3] [tid: 3728]     AgentID: /shared/SCHEDULED REPORT/Netbanking/15G_15H_Report [[

    ...Trying main Agent loop again.

]]

[2017-01-11T05:00:01.000+05:30] [OracleBISchedulerComponent] [ERROR:1] [] [] [ecid: 00id5k_UUcQ6mJ45JVXBif0000lO00007F,0:1:3] [tid: 3728] ... Sleeping for 8 seconds.

[2017-01-11T05:00:09.000+05:30] [OracleBISchedulerComponent] [ERROR:1] [] [] [ecid: 00id5k_UUcQ6mJ45JVXBif0000lO00007F,0:1:3] [tid: 3728]  [nQSError: 77006] Oracle BI Presentation Server Error: A fatal error occurred while processing the request. The server responded with: Inconsistent job ids.  Scheduler job id 105 points to ibot  [[

/shared/SCHEDULED REPORT/Netbanking/15G_15H_Report" with job id 444.  A catalog update has been

made without updating the scheduler tables.  The Administrator should consider deleting the old scheduler job.

Error Codes: C8TDLG6P

Location: saw.delivers.rpc.getMetadata, saw.rpc.server.responder, saw.rpc.server, saw.rpc.server.handleConnection, saw.rpc.server.dispatch, saw.threadpool.socketrpcserver, saw.threads

.

Error Codes: AGEGTYVF

Thanks,

Ramesh

Sagar Tippe

Hi,

Your error log clearly says,

1. you have entered wrong credentials / or access has been revoked for that particular user
2. Or else you have attempted to "Scan and Update Catalog Objects That Require Updates" which will change the object id's of some reports and other objects.

As OBIEE works with Object id's and not with the names. Why not try with creating same Agent again (from scratch) and run the same. I hope this will get resolved.

961614

Hi Sagar,

I have just rename that agent, it was trigger successful. All agent running from 2015 on wards but recently we are facing this issue randomly.

We have around 350+ agents. those agents trigger daily and monthly.

How to avoid this issue permanently?.

Thanks,

Ramesh

Joel

Have you tried recreating the scheduler job? There seems to be a mismatch between agent and NQ tables.

To test, create another Agent and see if it generates your report.

Sent from my iPhone

961614

Hi Joel,

Yes. It was working when we create new job. We are facing this issue only in old jobs but it was not continuously.

Thanks,

Ramesh.

Sagar Tippe

Hi,

The GUIDs are locked down to each user, yet there could be authentication issues causing strange behavior with GUID mismatches. The agents only make this issue more pronounced, and if you have GUIDs that are out of sync, then that could easily impact agents.

In this instance the scheduler tables are inconsistent (e.g. one job is pointing to another).

Two most common reasons for this issue are:

• Two unrelated schedulers (i.e. not a failover setup) are configured to use the SAME tables.
• User has broken table relationships (for reasons unknown)

Without a consistent relationship between these tables, there is nothing to prevent jobs from being created incorrectly. When two different Schedulers try to reuse the same structure, both get out of sync.

Before doing anything else the GUID issue will need to be fixed. (Google it, plenty of articles on this one)

Then proceed to fix the scheduler table.

1) In the Job Manager, delete the corrupt jobs (i.e. identify by a mismatch between related tables). The schema set up script should help identify the expected contents of the tables.

Using errors like “The Administrator should consider deleting the old scheduler job” should help to identify the mismatches.

If, and only if, this does not help:

2) The worst-case-scenario work-around is to create a new scheduler schema for each environment, as applicable, pointing the scheduler(s) to the new database schema and recreate the scheduling information in the Delivers interface. The actual agents will still be present, but none of them will have any scheduling details. Each will need to be re-entered/saved. You can refer to the old schema to determine the previous settings and once completed, you can drop the scheduler schema.

Hope this helps.

961614

Hi Sagar,

How can I get script?.

"The schema set up script should help identify the expected contents of the tables."

Thanks,

Ramesh

Sagar Tippe

Ramesh,

Go to job manager and delete the failed jobs manually, no need of script.

DentsplySirona

That is fine unless you have a lot of jobs with some scheduled to run once a week, month or year.  How do you fix this for jobs that are to run in the future without know what they are.  We just upgraded from 10g to 12c and are having this issue.  I am able to do the manual process to fix the ones we know about, but how do I fix the future ones before they happen.