DBIF_REPO_SQL_ERROR - ORA-03114: not connected to ORACLE

[Former Member]

Hello,

In our SAP ERP system, we are having database connection lost with application servers frequently. This is happening weekly 3 times.

During disconnection of database, all the background job getting cancelled and receiving the dump

*************************

Runtime Errors	DBIF_REPO_SQL_ERROR
Date and Time	18.07.2016 09:43:04

Short text

SQL error 3114 occurred when accessing program

What can you do?

Note which actions and input led to the error.

	For further help in handling the problem, contact your SAP administrator
	.

	You can use the ABAP dump analysis transaction ST22 to view and manage
	termination messages, in particular for long term reference.

ow to correct the error

   Database error text........: "ORA-03114: not connected to ORACLE"

   Triggering SQL statement...: "CX_SY_SQL_ERROR===============CP "

   Internal call code.........: "[REPO/*/43/LOAD/CX_SY_SQL_ERROR===============CP

"**************************************************

We could not able to find any connection lost logs in database level.

Please help me to identify the cause and solving the issue.

Thanks and Regards,

Anand

[Former Member · Accepted solution]

Hi All,

Thanks for the update.

We are taking tcpdump in one server and analyze the logs from network team.

And also we migrated another system database server to another EsX host and monitoring the same.

I will update you based on the result.

Thanks,

Anand

[Former Member]

Hello All,

Thanks for your feedback. Now we are not having the network interrupt in the database server.

We are not sure what is the root cause of this issue and how it got solved.

Regards,

Anand

[Johan_sapbasis]

Hi,

Please paste a section of your alertlog here and listener log files here from the time of one of the errors.

ORA-03114: not connected to ORACLE

Also attached a work process trace file.

Troubleshooting wise have you checked your application servers dev_rd for any errors during startup i.e. network?

Can all servers ping each other dns/hostname without issues. As mentioned by sriram check maybe that they are all in the same subnet.

Also when did this issue start i.e. was there a recent change in the environment server/sap/network before issues started occurring?

Kind Regards,

Johan

[Former Member]

Hi Anand,

Can you share some details regarding your SAP Architecture. OS/DB release and how are the instances (SAP and DB) distributed?

You may also want to check with the infrastructure team on the network usage report for the duration you get these disconnects in your system.

Thanks,

Manas

[Former Member]

Hi Manas,

OS Version - SuSE release 11.2

DB - Oracle 11.2.0.3

SAP Version : EhP4 EHP4 FOR ERP 6.0_NW701 SP16

System configured with distributed environment. Database and Application installed in separate instance.

We requested our network team to check the issue. But we got update saying there was no drop during that time.

Can you please let me know what exactly need to check with network colleagues.

Thanks,

Anand

[Former Member]

Hi Anand,

You can start with setting up network traces using note 500235 - Network Diagnosis with NIPING.

Also check Troubleshooting Common Errors and provide the alert.log for checking further.

Thanks,

Manas

[Sriram2009]

Hi Anand.

1. Could you check the over all free disk space in Oracle DB.

2. Could you share the full dump as text attachment. and also you can find the disconnection reason in SM21 log.

3. Your Application and DB instance are same local LAN or other location ? Check this SAP KBA

2064591 - work processes reconnect to Oracle database with error ORA-03135

BR

SS

[Former Member]

Hi Manas,

We established niping between application and database servers. We observed the below logs in application side during the database connection lost.

The received data pac number changing during the interrupt.

It means we have network drop between application and database servers?

NiWait: sleep (1000ms) ...

Tue Jul 19 09:23:58 2016

NiIWrite: hdl 1 sent data (wrt=10000,pac=1,MESG_IO)

NiIRead: hdl 1 recv would block (errno=EAGAIN)

NiIRead: hdl 1 received data (rcd=10000,pac=2,MESG_IO)

NiWait: sleep (1000ms) ...

Tue Jul 19 09:23:59 2016

NiIWrite: hdl 1 sent data (wrt=10000,pac=1,MESG_IO)

NiIRead: hdl 1 recv would block (errno=EAGAIN)

NiIRead: hdl 1 received data (rcd=10000,pac=3,MESG_IO)

NiWait: sleep (1000ms) ...

Tue Jul 19 09:24:02 2016

NiIWrite: hdl 1 sent data (wrt=10000,pac=1,MESG_IO)

NiIRead: hdl 1 recv would block (errno=EAGAIN)

NiIRead: hdl 1 recv would block (errno=EAGAIN)

NiIRead: hdl 1 received data (rcd=10000,pac=3,MESG_IO)

Thanks and Regards,

Anand

[Former Member]

Hi Sriram,

Please find the attached dump.

Our application and database servers placed in same data center and there is no firewall inside data center.

Thanks,

Anand

dump.txt.zip

[Sriram2009]

Hi Anand.

1. Have you got any error message during the dump & DB disconnection in transaction code SM21 log?

you kernel release, level and DB client are same in CI & App ?

2. From the Dump while executing the Z report (ZPP_MET_REPORTS) DB got disconnected. You can check the over all free disk space in application server, DB & log folders.

3. You can check with Network admin team to check the connectivity from App to DB.

BR

SS

[Former Member]

Hi Anand,

Check the sap note

2152158 - Intermittent Oracle connectivity issue

Also, as the call is made from a Z report please involve ABAPers to debug the case.

Regards,

Prithviraj