"Program terminated in remote system NONE: Logon failed" when adding a new table

[chri_schmitt]

Hello,

I set up a replication flow from a SAP ERP 6.0 EHP7 with SAP ASE 16.0 source to SAP HANA 1.0 rev 82 target
I am using a standalone SLT system NetWeaver 7.0 with DMIS 2011_1_731 (with SP 1 to 7).
I have 100 tables to replicate.

I succeeded to set up replication for 57 tables. These tables are replicating properly.
Whenever I try to add a new one (with LTRC transaction, Data Provisioning -> Start Replication), the new table is marked as 'Failed' after a little while.

When I press 'Show Error Log' button I got an obscure error message:

"Program terminated in remote system NONE: Logon failed "

I do not understand this message. I checked both on SAP ASE source and SAP HANA target. I am still able to connect against both source and target.

Can you please tell how to troubleshoot this error.

Thanks in advance,

Christian

[chri_schmitt]

First thank you for answering my questions. I really appreciate your answers.

I rechecked the documentation.

"Application Operations Guide SAP Landscape Transformation Replication Server Document Version: 2.3 – 2014-07-08"

Page 29 - 30

3.5.2.2 Data Transfer Jobs
This section explains the relationship between the number of data transfer jobs and the number of available background work process.
Data transfer and data transformation processing on SLT server system is accomplished by the background work processes of the underlying SAP NetWeaver ABAP application server. Each job occupies 1 background work process in the SAP LT Replication Server system. For each configuration, the parameter Data Transfer Jobs restricts the maximum number of data load job for each mass transfer ID (MT_ID). In total, a mass transfer ID (MT_ID) requires at least 4 background jobs to be available:
 One monitoring job (master job)
 One master controller job
 At least one data load job
 One additional job either for the migration objects definition, access plan calculation or to change configuration settings in the Configuration & Monitoring Dashboard
Example
If you set the parameter Data Transfer Jobs to 04 in a configuration “SCHEMA1”, a mass transfer ID 001 is assigned. As a result, the following jobs should be in the system:
1 Master controller job: /1LT/IUC_REP_CNTR_001
 At most 4 parallel jobs for MT_ID 001: /1LT/IUC_LOAD_MT_001_001/~002/~003/~004
When configuring your data load or replication scenario, consider the following:
 Do not define more data transfer jobs than the number of available application server background work processes. If all available background work processes are already occupied by jobs, any other job will have to wait until a free work process becomes available. This can lead to long wait times until a new activity (for example creating triggers) can start, and can also result in significantly increased latency times for data replication.
 The number of dialog work processes in the source system corresponds 1:1 with the number of data transfer jobs in the SAP LT Replication Server system.
 Besides the work processes allocated by the data transfer jobs you need to provide additional available work processes for controller and monitoring jobs, the migration objects definition, access plan calculation or to perform configuration changes, and so on.
Sizing for SAP LT Replication Server involves determining how many work processes are required to perform the initial load of data into the target system within an acceptable timeframe, and accomplish the change capturing and the transfer of data changes to the target system within expected latency times.
Ensure that you add enough additional work processes to allow other required SAP LT Replication Server jobs to run.
Finally, you map the number of required application server work processes to their system resource consumption (CPU, memory, disc space) using the formulas provided by the SLT Sizing Guide.
With the simple formula below, you can calculate the number of required application server work processes (WPs) on the SLT Server for each active SLT configuration.
The number of required work processes can be determined by adding
 The Number of required data transfer jobs ,
 plus one background work process for Central Master (Monitoring) Job (only one per system!),
 plus one background work process for Master Controller Job,
 plus 3-5 additional empty background work processes (recommended per configuration),
 plus approx. 3 dialog work processes (recommended for each configuration).
Note: A lack of available free application server work processes can negatively affect the data load or data replication processes.

To summarize everything, the number of 'Data Transfer Jobs' must be set depending of the number of source tables, it is not the actual number of tables.

Assume that for my 100 tables I use 10 'Data Transfer Jobs' :

- The number of work processes on the SLT server would be 20. I took the simple formula of the documentation:
10 data transfer jobs ,
+ 1 background work process for Central Master (Monitoring) Job (only one per system!),
+ 1 background work process for Master Controller Job,
+ 5 additional empty background work processes (recommended per configuration),
+ 3 dialog work processes (recommended for each configuration).

- The number of dialog processes on the source server would be 10 ( equal to the number of 'Data Transfer Jobs')

Am I correct ?

Regards,

Christian

[srinivasan_vinayagam]

Hi Christian,

Definitely all SAP servers having above dialog and background processes what you calculated.

You should confirm the background work processes before replication. All customers should maintain dialog processes more numbers.

Regards,

V Srinivasan

[srinivasan_vinayagam]

Hi Christian,

Please check this thread: -

Regards,

V Srinivasan

[chri_schmitt]

Hello Srinivasan,

Thank you for your answer. I read the thead you mentioned, but did not find sizing information for SLT.

If I want to replicate 100 tables, how must I set following parameters:

- On source ERP NetWeaver server:

rdisp/wp_no_dia   = ?

- On SLT Netweaver server:

Data Transfer Jobs =?

rdisp/wp_no_btc    = ?

Regards,

Christian

[srinivasan_vinayagam]

Hi Christian,

rdisp/wp_no_dia   = ?

rdisp/wp_no_btc    = ?

Above parameters are number work processes.

How this related to 100 Tables?

Regards,

V Srinivasan

[srinivasan_vinayagam]

Hi Christian,

This link help for and fix your issue.

Thanks & Regards,

V Srinivasan

[chri_schmitt]

I agree. This is a smart question.

I am convinced that this issue is related to configuration sizing, because everything works below a certain level of replicated tables.

I read following document:

"Application Operations Guide SAP Landscape Transformation Replication Server Document Version: 2.3 – 2014-07-08"

Page 60

"

The source system and SLT system should be configured to have appropriate number of background jobs and dialog work processes to ensure parallel execution and full allocation of the hardware configuration which is sized for the expected replications. An insufficient number of work processes results in a sequential execution of the replication and therefore in much longer execution times. On the other hand, an insufficient number of available work processes in the source system may impact the performance of business applications running there.
For each SLT configuration, the parameter Data Transfer Jobs restricts the maximum number of data load jobs, which can be started for one mass transfer ID (MT_ID).
In the source system, the number of available DIA work processes, which are reserved for SLT replication should be equal to the number of configured Data Transfer Jobs on SLT side. In addition free work processes are required when creating new logging tables and database triggers, as well as creating runtime objects and calculating access plans."

Can someone explain me the purpose of background/dialog processes with SLT replication and how to tune these parameters.

Regards,

Christian

[chri_schmitt]

Thank you very much for the link.  SLT: Mass Tables for Replication to HANA

I was not aware that HANA Studio Modeler can be used for seting up SLT replication.

I checked on my HANA Studio. The Data Provisioning is not configured.

I cannot use it. But, it is helpful to know that we can use it.

I use the "The Better Way". The rest of this section shows how to mass declare table replication.

This is very very useful information for me as I have a lot of tables to replicate.

Thank you again Srinivasan for this very valuable information.

But my question remains, how to tune dialog/background processes for SLT Replication ?

Regards,

Christian

[srinivasan_vinayagam]

Hi Christian,

The Number of Work Processor calculated by SAP system Hardware related sources.

For Total Number of Work Processor :

Determining the Number of Work Processes - SAP Memory Management (BC-CST-MM) - SAP Library

No.of dialog and No of Background determine by Physical memory

No.of Dialog WPs=RAM Size(MB)/256 & No.of Background Wps=RAM Size(MB)/768 - Approx. not sure

Regards,

V Srinivasan

[Former Member]

Hi,

No.of Dialog WPs=RAM Size(MB)/256 & No.of Background Wps=RAM Size(MB)/768..

Even if it is an approximate but just a quick query...how did you get this figure 256 & 768?Is it there in ay SAP document or just from your experience you are suggesting?

Regards,

Joydeep.

[srinivasan_vinayagam]

Hi,

This is from certification Materials.. TADM.

Thats why i have mentioned not sure.

[Former Member]

Hi Christian,

Please check the RFC Connection(Authorisation as well as Connection test) between your SLT & Source system.Secondly the DBCO conn between HANA & SLT.

Are you using a separate Tablespace for the logging tables of the source?

Regards,

Joydeep

[chri_schmitt]

Hello Joydeep

Thank you for your answer.

Q: Are you using a separate Tablespace for the logging tables of the source?

A: The source system is SYBASE ASE. Tablespace do not exist. Why this question?

Q: Please check the RFC Connection(Authorisation as well as Connection test) between your SLT & Source system.Secondly the DBCO conn between HANA & SLT.

A: I think that the authorizations are correct, otherwise I would not be able to set up the replication for the 57 first tables.

I read following document:

"Application Operations Guide SAP Landscape Transformation Replication Server Document Version: 2.3 – 2014-07-08"

Page 60

"The source system and SLT system should be configured to have appropriate number of background jobs and dialog work processes to ensure parallel execution and full allocation of the hardware configuration which is sized for the expected replications. An insufficient number of work processes results in a sequential execution of the replication and therefore in much longer execution times. On the other hand, an insufficient number of available work processes in the source system may impact the performance of business applications running there.
For each SLT configuration, the parameter Data Transfer Jobs restricts the maximum number of data load jobs, which can be started for one mass transfer ID (MT_ID).
In the source system, the number of available DIA work processes, which are reserved for SLT replication should be equal to the number of configured Data Transfer Jobs on SLT side. In addition free work processes are required when creating new logging tables and database triggers, as well as creating runtime objects and calculating access plans."

In my case, I have 100 tables to replicate, does it mean that I need at minimum fro replication purposes:

- On source ERP NetWeaver server:

rdisp/wp_no_dia   = 100

- On SLT Netweaver server:

Data Transfer Jobs =100

rdisp/wp_no_btc    = 100

- On target SAP HANA server:

100  user sessions defined

Thanks for your support,

Regards

Christian

[Saritha_K]

Hi Christian,

Can you try viewing SM21 and ST22 t-codes for any aborts in SLT system?

Check in SM50 for the number of available free jobs..

from T-code LTRC-> Load statistics- how many tables are in progress and finished state?

If the list of tables that you gave initially are replicated completely, please select few tables for replication this time.

i.e you can start replication a set of tables in batches may be 10 at one go or something rather than defining 100 jobs and user sessions.

Regards,

Saritha K

[chri_schmitt]

Hello Saritha,

I reconfigured both source and SLT system because they were not stable.

The SM21 and ST22 do not hold error message now.

I ran the SM50 transaction on both source and standalone SLT server. Please see below the output.

SLT server:

Time    18.11.2014, 09:45:27   
Total Number of Processes    63   
Dialog    10 / 9 (Total/Free) , Average Load = ( 0,024 / 0,005 / 0,007)   
Update    1 / 1 (Total/Free)   
Background    50 / 7 (Total/Free) , Average Load = ( 42,993 / 40,127 / 38,216)   
Spool    1 / 1 (Total/Free)   
Update Task II    1 / 1 (Total/Free)   
Configurable    63   
Preemptions    3 

ERP source:

Time    18.11.2014, 09:46:11   
Total Number of Processes    24   
Dialog    15 / 7 (Total/Free) , Average Load = ( 8,326 / 8,014 / 7,600)   
Update    1 / 1 (Total/Free)   
Background    6 / 6 (Total/Free) , Average Load = ( 0,006 / 0,002 / 0,001)   
Spool    1 / 1 (Total/Free)   
Update Task II    1 / 1 (Total/Free)   
Configurable    19   
Blockade Handling    2.208   
Preemptions    5   

I checked the HANA target:
indexserver.ini -> session -> maximum_connections = 65536
indexserver.ini -> session -> maximum_external_connections = 65536

I reduced my ambitions to replicate 42 tables.

The "Data Transfer Jobs" is now set to 50.

I checked LTRC -> Load statistics. All the tables are in finished state.

The system is stable with the current amount of tables.

Nevertheless, my question remains. Assume that I want to replicate 100 tables, how must I set following parameters:

- On source ERP NetWeaver server:

rdisp/wp_no_dia   = ?

- On SLT Netweaver server:

Data Transfer Jobs =?

rdisp/wp_no_btc    = ?

Regards Christian