İs there a general problem with Oracle 11.2.0.4 on windows2012. ora-3113,ora-3114 and ora 3135

[Former Member]

Hi

Last two months we have installed different SAP products for different customers on Oracle 11.2.0.4 with Windows 2012 R2 but all of these installation giving ora-3113,ora3114 and ora-3135 error.Some of installation completed successfully but the installed system is not running properly besides that when we opened the incident SAP sides always offer to use niping to send the log to SAP and they said that "There is a  network problem".Special thanks to SAP I can also read the logs !!! Yes it's has written in the log "there is a network problem".There is a nework problem but there is not a solution yet.To solve problem i have changed INBOUND_CONNECT_TIMEOUT and some sqlnet.parameters also changed tcp parameter from windows registry described in 534729 and 1911531 by the advice of SAP consultants. I think it is not specific problem , it can be kind of noncompatibility between windows 2012 R2 and Oracle 11.2.0.4

Do you have any comments and experiences about the oracle errors mention above?

Regards...

[ACE-SAP]

Hello

Are these systems virtualized ?

Does the parameter INBOUND_CONNECT_TIMEOUT  fix your issue ?

Are there network errors reported by other tools on these systems ?

Did you performed niping test as requested by support ?

It's surely not an incompatibility between 11.2.0.4 & Win 2012R2 as both SAP (check the PAM) & Oracle (http://www.oracle.com/technetwork/database/windows/tech-info/sod-oracle-db-win81-win2012r2-2088197.p...) has validated them.

Have you tried deactivating TCP Chimney Offloading & RSS (Receive-Side Scaling) ?

These parameter could cause network problems on older Windows versions.

Its easy to change, just need a reboot to be active.

You can check if they are active with:

netsh interface tcp show global

and disable it with:

netsh interface tcp set global chimney=disabled

netsh interface tcp set global rss=disabled

You could also try changing "Receive Window Auto-Tuning Level"

netsh int tcp set global autotuninglevel=disabled

Regards

1519484 - How to analyze network disconnections shown in system log (BC transaction SM21)

Also, we have found a lot of issues with some new features as the Scalable Networking Pack aka. SNP (TCP Chimney Offload feature, RSS, and NetDMA). In particular, we always recommend to disable the "TCP Chimney Offload" feature option on your NIC. In order to do so, you can run from a command prompt “netsh int ip set chimney DISABLED”; run “netsh int ip show chimney” in order to know its current status. Then, reboot the system (it is mandatory!). Please, check points 6 to 8 in the “Solution” section of note 392892 (MS KBA 942861), also if your database is other than MS SQL Server.

http://blog.wapnet.nl/2014/04/windows-2012-r2-unstable-internet-connection/

[Former Member]

Hi Yves

Yes all of these system are running on VMware esxi

Changing INBOUND_CONNECT_TIMEOUT parameters hasn't been effected.

For the other programs  it  hasn't been occured.

I have used niping tool but i don't know how to use niping tool efficently further more i also believe niping tool is unnessary.6 months ago we have updated a SAP ERP system's database which is sybase.When the update process completed system required restart.After restart we have recieved dump error message and it is also related with communication error between db and client.When we opened incident on support.sap.com ,guess what sap said? Please use nipping tool as descriebe in note bla bla bla.Nothing hadn't been changed on that system just the database program patched.Why did they required to use niping i doesn't understand.I don't believe niping is a effective tool.

Last month i have installed 4 system also my colleague installed a system.All of system mentioned are ABAP  system with oracle 11.2.0.4 DB running on Windows 2012 r2.we have recieved similar error.One of our customers firstly wanted to install system MS-SQL and then we installed the system as they like and all went very well afterwards customer change his decision and they wanted a system with oracle running on windows 2012 r2 so that we uninstalled sap system and MS-SQL server 2012.When we began to install system we faced the oracle problem(ora-3113,ora3114,ora3135).For your kind attention it is completely same system but Oracle 11.2.0.4 had been trouble.On same system MS-SQL runs properly but oracle not

I have also applied your recommendation but the result is same it is still giving error message

Any idea about VMware and Oracle compatibility?Is the problem related with it?

Special thanks Moshe and you

[ACE-SAP]

Hi

I was not able to find any information about problems on 2012R2 with 11.2.0.4.

So you did get the same problem with Sybase, or you did get an error after migrating that system to Oracle ?

On the systems where you get that problem is the DB running on the same host as the SAP instance ?

Did you apply the latest bundle patch (7) on these systems ?

For niping you need to start it on the server where the DB is running using the port dedicated to Oracle

niping -s -I 0 -V 2 -S 1521 -T c:\temp\niping.log

-s :       server mode

-I 0 :     no automatic shutdown

-S :       service-name or port-number

-V 0...3  Trace level

and you launch it in client mode on a SAP instance host

niping -c -H <db_server_hostname> -S 1251 -B 131072 -L 20 -D 500 -V 2 -T c:\niping.log

-B <buffersize> (default 1000 bytes) determines the size of the data packets.

-L <loops> is the number of packets sent (default 10).

-D <delay> delay between requests (in milliseconds, defalut 0).

Regards

500235 - Network Diagnosis with NIPING

1100926 - FAQ: Network performance

[Former Member]

Hi

I am sorry for misunderstanding

The problem about sybase completely disparate.I just wanted to give an example to share my idea about niping tool usage

The system and database are on the same server.

Yes I have also applied the bundle patch and the others.

I have atteched the niping tool's log

Regards

niping_client.txt.zip

niping_db.txt.zip

[ACE-SAP]

Hello,

I did forget to tell you that you should first stop Oracle listener in order to be able to perform niping test on the port it is using (that is by default 1251).

But tcp tests are useless as your SAP system is running on the same host as the Oracle DB.

In this configuration the communication between SAP workprocess & the DB goes through IPC (shared memory) and not SQL*Net.

You can check if IPC has been explicitly disabled in the lisnter.ora (AUTOMATIC_IPC=OFF) but this is not the default configuration.

You should activate trace on the listener, either in the listener config file or through lsnrctl

You will need at first to disable security option ADMIN_RESTRICTIONS_LISTENER = off in listener.ora conf file

Then from lsnrctl prompt, restart listener and activate trace

  set trc_level 4

Check where the trace is generated

show trc_directory

show trc_file 

Do not forget to revert changes made in listener.ora conf file & to stop the trace.

      set trc_level OFF

Regards

562403 - FAQ: Oracle Net

[Former Member]

Hi Musa,

I have also came across a very similar issue last week, while installing SAP application server on Windows 2012 R2 while using Oracle 11.2.0.4.

I have connection problems with DB (errors 3113,3106). We don't have any issues with other application servers that are not Windows 2012 R2.

With 2012 R2 server we have also a problem with "inserts" that hang on the DB side with wait type "more data from client". The sessions never end on DB side and work processes hang on SAP side.

Currently we have no solution to these issues.

Regards,

Moshe