RFC Server Groups affecting RFC performance

Hi SDN performance experts
I am investigating an issue with RFC performance within our ECC5 system. ST03N shows an average total RFC response time across all instances has grown from a monthly average of 352ms in March to 1500ms in October. However, CPU and Database times have not risen very much during this time, indicating that we are 'losing' time elsewhere.
I broke down my investigation to look at each instance in turn and what I found was that half the instances had an average monthly RFC time that had been stable since March, whilst the remaining instances showed times that had increased greatly in May and June. Two of these instances had shown an increase from 0.5secs to nearly 20secs in three months!!
I dug a little deeper and found that a RFC server group (RZ12) had been created for the instances that are showing poor performance around May or June. The instances NOT included in this RFC server group are the ones that continue to perform well. I therefore assume that the problems are caused by this RFC server group.
The group has the following attributes (within RZ12):
Activated - 1
Max requests in queue - 5
Max no of logons - 90
Max disp of own logon - 25
Max no of WPs used - 75
Min no of free WPs - 1
Max no of comm entries - 90
Max wait time - 15
Can anyone help me solve this problem by advising how the RFC group should be set up to improve performance.
Many thanks,
Arwel Owen,
SAP BASIS Manager,
Princes Ltd,
Liverpool, UK.

Hi Nuno,
We configured RFCs to use application instances instead of the central instance because we too had problems with RFCs locking up the central instance many years ago. We took the decision then not to run RFC-based interfaces from the central instance, but to balance them across the application instances instead.
Each application instance has 15 DIA processes and the system barely ever reaches any near full capacity (where all DIA processes are in use), so I think we have enough processes configured.
We are not experiencing many buffer swaps. The Export/Import buffer has very similar number of swaps across all instances - it does not show more swaps for instance 01 over instance 00 (see my previous post for more information on how our instances are configured).
Can you please explain what you mean by configuring SMLG instead of RZ12 for RFCs?
Many thanks,
Arwel.

Similar Messages

Startrfc & RFC Server Groups

Good day.
I was curious if there was a command line tool, similar to startrfc, that would leverage the RFC Server Groups (RZ12) as opposed to the Logon Groups (SMLG).
We currently have 2000 iDocs to post at once and are bombarding a single APP server since the logon load balancing algorithm is not rapid enough. Since the RFC Server Groups are designed for this exact purpose, I find it silly that startrfc would not leverage them.
Any help or ideas would be greatly appreciated.
Thank you,
Charles
DBA & BASIS Admin

Hi Charles,
We are developing an interface between SAP and an external system.
Estimated load for inbound documents is about 5000 iDocs per hour.
We plan to use "startrfc" command to load IDocs to SAP.
We are also interested in using RFC Server Groups (RZ12).
Have you found a WorkAround ?
Best regards
Jean Christophe

SOAP vs RFC - performance and pros and cons

Hi,
I have been playing around with the SAP .NET Connector 2.0 between an ASP.NET 1.1 application and SAP passing in either SOAP or RFC connection strings to the proxy classes as a proof of concept for some integration work required for our existing customer website and teh SAP system replacing our legacy applications.
Other than issues relating to the ports that need to be opened in Firewalls between application environments can anyone provide me with any information or point me in the direction of any documentation regarding the performance differences between and pros and cons of using these two protocols?
The impression I get is that RFC performs better than SOAP due to the XML conversion overhead inherent in using SOAP calls.
I know that the ESOA direction of SAP means that SOAP is the desired protocal in the long run but is there any reason why one should be used in preference to the other in the scenario menttioned above?
Sorry if there is an obvious answer and thank you in advance for any help.
Andy

We did a lot of load tests to find answers on how the protocol can have impact on the performance. With our test systems RFC was about 3 times faster than SOAP and the network traffic was doubled with SOAP (approx.). And it seems to me that RFC offers more parallel processes (for multithreading application), but I'm not hundred percent sure on that point.
We used rel. 640 with Web AS, Client was .NET Connector and ERPConnect with .NET 1.1 and .NET 2.0.
Krishtan

What could mean "server group"

What could mean server group. I have to specify when running a program in background(probably this is not a group in "smlg"as this partains to dialog processing

I think you're refering to SAP RFC server group
Similar to load distribution of user logins with logon groups, there is a mechanism for load distribution of RFCs
This is needed to achieve parallelization to achieve higher performance
To avoid overloading a dialog instance with too many simultaneous RFCs, it uses the resources of all existing dialog instances optimally. Hence you refine RFC server groups.
Use RZ12 to maintain RFC groups.
Create a group & assign single/multiple instances to this group
Regards
Puneeth

How to setup "Server Groups" in RSPC (process chains)?

I'm a BASIS person working with our BI group on setting up their loads to only run on the application servers. If you don't define a host in RSPC, it will always run on the database server. We don't want to specific one host in case that server has issues. How can we setup the "server group" so we can use all our application servers for these process chains. It doesn't seem to use the SMLG logon groups we defined already.
Thanks in advance!

I have setup the load balancing in the RFC, but for these jobs that are submitted via RSPC, they don't look at the SMLG parameters, they just always pick the database server. We can specify one host (other than the database server) in RSPC similar to SM36, but we were trying not to do that so in case that server is down, these jobs wouldn't fail.
I thought we could setup a "server group" and assign our 3 application server to them that way they would never fail if one is unavailable and it would keep them off the database server.

BODI 11.7.3.0 could not connect to a SAP server group

We are currently using Business Objects Data Integrator 11.7.3.0. and we trying to connect to SAP server group instead of specific server but getting connection error
We are setting this up because BODI is not ulitizing the Load Balancing feature of our SAP system. Because the SAP application server address is hard coded in BODI (Datastore), BODI is just using the same SAP server every run.
The question now is: Is the latest version of BODI can connect to a SAP server group?

1. Yes, 12.2 finally suports the rfc.ini file, one reason was for the load balancing.
2. Yes, you are right, one datastore is bound to one application server in 11.7. However, this one apllication server could submit the ABAP as a spool job rather than executing it in dialog mode and then you have your load balancing again. Checkout the datastore property "execute in background".

Server group management

Hi,
When setting up an architecture where servers should be accessible only to a subset of the platform's users, e.g. geographically dispersed installations, SAP has always said to use server groups and assign the content (i.e. the reports) to the corresponding server group.
Example:
Reporting databases are deployed in Belgium and France
2 hosts containing processing servers are installed in each of those countries
Reports are restricted to one of the two server groups
While this setup works fine for existing reports, what do you do with ad-hoc reports? E.g. a user creates a new Web Intelligence report against a universe that pulls data from the database in Belgium. As far as I can tell, there's no way that you can make that report use the processing server in Belgium. Instead, it will randomnly use a processing server that's available to handle its request.
Also, using server groups this way requires your users to set the processing server after they save the report for the first time, or create custom code based on the SDK that will assign the server group.
I've run into the same problem with deployments where there's a requirement to have processing servers dedicated to departments. As far as I can tell, there's no way to do this as there's no link between user and server (i.e. it's not possible to define which users or user groups have access to a server group).
I was hoping the multi-tenant feature in BI4 would provide some way around this, but it seems that's not how it works.
Any input regarding this type of setup is very welcome! Thanks.
Kind regards,
Kristof Speeckaert

Hi,
Thanks for your input.
I tried this in the past, but just to be sure, I did the following setup again (BI4 SP05):
Created two servers groups
Server group 1:
WebI Processing Server 1
Connectivity Adaptive Processing Server
WebI Adaptive Processing Server
Server group 2:
WebI Processing Server 2
Connectivity Adaptive Processing Server
WebI Adaptive Processing Server
Created one user group with one user in it
Denied access to server group 1 for the test group
The user group has full application and content rights
Scenario:
Logged on using the BI Launch pad
Opened and refreshed a sample WebI document
Brought the Webi Processing server in group 2 downOnly the WebI processing server in group 1 is up and running and the user group has no access to this server group.
Logged out and in again
Opened and refreshed the same sample document
Based on your suggestion, I shouldn't have been able to open nor refresh the sample document in step 5, however it worked without a problem.
This would mean that granting/denying users access to servers/server groups does not prevent them from using the server, in this case for opening and refreshing a WebI document.
Here's the relevant paragraph regarding server group access rights from the Administrator Guide:
You can use rights to grant people access to servers and server groups, allowing them to perform tasks such as starting and stopping servers.
If your server team needs to perform regular server maintenance tasks that require them to shut down and start up servers, you need to grant them rights to the servers. You may also want to delegate SAP BusinessObjects Business Intelligence platform server administration tasks to other people. Or you may want different groups within your organization to have control over their own server management.
Judging from this information, assigning rights to server groups should be used from administrative access to the servers, not to decide who get to use the server from an end-user perspective. The latter should be done by assigning server groups to documents/reports. However, this leaves us with the same problem when dealing with ad-hoc reports.
Kind regards,
Kristof Speeckaert

BOXI3.1 How to dedicate server groups to select Webi reports?

We have a select number of Webi reports that we would like to always run on a specific server group which consists of a number of WebIntelligenceProcessingServer(s). I can get this to work fine by 1) setting up the server group 2) Specifying that server group in the "Default Settings" of a Webi report using the CMC 3) scheduling the report and viewing the latest instance. However I can find no way of stopping all other Webi reports in the system from running on that same server group.
Anyone know how to set a server group up so that it only processes specific reports and not ever report in the entire BO XI repository?
Thanks,
George

I have set up a little test on my BOE 3.1 SP3 system and have found the following:
2 server group: SDKNO and SDKYES
4 WEBI Server (2 ADP and 2 WPS)
Have given SDKNO a set of WEBI servers and have given SDKYES the other WEBI servers
Created 2 user group SDKNO and SDKYES
Have set SDKNO group to have NO acess to SDKYES and full control to SDKNO have set SDKYES group with full control to SDKYES and No Access to SDKNO
Have created and schedlued webi reports as a user that should have NO access to the SDKNO server group and reports will still use the SDKNO servers.
I have gone one step further and set security directly to the servers as opposed to the server group and still a member of the "SDKNO" group can use the "SDKYES" servers.
Apperantly there does not seem to be a way to isolate servers based on group membership.
What does this mean to you?
Basically, if you want to use serve groups and you want to ensure that no "other reports" use any available server then all reports MUST be assigen to a only use a specific server group as, if this is not set, the "use first avaialble" option will come into affect and trump any server group security settings that have been set up.

Server 2012 R2 slow performance over all

This is a DELL PowerEdge R820 / 256GB RAM / 4TB onboard storage configured as Remote Desktop Services Host /TS... running QuickBooks, MS Office 2010 STD, Symantec EndPoint Protection (basic antivirus installation ONLY)... 1 month old server... SUPER fast,
super powerfull.... Part of an AD in SBS 2008 Premium... single NIC card 4 ports / 3 ports disabled. single IP... no VLANS... server is able to resolve ANY pc, DNS record with no problems... able to resolve DC by name, able to receive GP, able to update, able
to do everything that I can think of.... EXCEPT... anything I open, word, excel, QuickBooks, IE.... takes forever to open... after the application is open is fast but to open anything takes 3-4 minutes.... open ticket with Symantec all looks good, created
lots of exceptions for antivirus/real time scanning, open ticket with QuickBooks: files look good application was removed and reinstalled to be sure all was done correctly... check the binding order for the disabled NIC; the active one is TOP option, no errors
at all in event viewers for system, application, setup, no errors at all in DELL management tool, no hard drive errors, no controllers errors.... this server is replacing an old dell poweredge 2008 STD with 24GB RAM... the old DELL opens the same applications
way faster than the new one... same Quicken version; old server opens Quicken in seconds, new one 3-4 minutes..... Real time monitoring NEVER goes above 1% for CPU and 3% for memory utilization.... one more thing... removed antivirus 100% restart... same performance
without antivirus.....
Any ideas will be great as of how to troubleshoot the slow performance....
Thank you!

Hi,
As Sam suggested, please check if there any issue occurred in hard drive.
On current situation, please also refer to following steps and troubleshoot, then check if we can find more
clues.
Please check if you have installed all necessary updates for the Windows Server 2012 R2.
Please
perform a clean boot to check if there has software conflicts.
Please use Resource Monitor to troubleshoot and check if we can find some more details.
Using Resource Monitor to Troubleshoot Windows Performance Issues Part 1
Using
Resource Monitor to Troubleshoot Windows Performance Issues Part 2
If any update, please feel free to let me know.
Hope this helps.
Best regards,
Justin Gu

Executing a Job via Web Services using a Server Group?

When executing a job from Designer or from the Management Console you have the option to select a specific job server or a server group.
But when executing a job via Web Services it uses the first job server with which the repository was associated (i.e. the first job server listed in AL_MACHINE_INFO).
Is there a way to get the job to execute using a server group instead simply choosing the first job server in the list?
Note: We are using FIM to execute the jobs via web services and don't have access to additional parameters.

Joe
Here is the where you can change the server group/job server.
- After you add the Batch/real Time job as a webservice.
- Go to the Webservice Batch Job Attribute and from the drop down choose the "Enable job Attributes" and hit apply.
- After this if you use the WSDL URL http://<Data Services Web Server>:<Web server Port>/DataServices/servlet/webservices?ver=2.0&wsdlxml
- You will see the job attributes that you can change based on the allowed parameters mentioned in the documentation.
The batch job or real time job is published without the job parameters by default. Once you enable the job attributes, remove and re publish the batch job as a web service. The application making the call to this webservice should be able to read the WSDL generated by Data services and pickup all the input and output parameters of the method.
Refer to the Data Services Integrator guide (Chapter 3 : pages 25 - 33 ) to get the exact parameter values permitted via Data Services jobs called via web services.
Hope this helps!
Thanks & Regards
Tiji

Bit Locker Implementation in Windows 8.1 machine using Windows server 2008 r2 server group policy.

is it possible to enable the bit locker only for windows 8.1 machines through windows 2008 r2 server group policy ?
Thanx and Regards,
Shanif

Hi Shanif,
Yes, we can do this.
Regarding how to enable Bitlocker via group policy, the following article can be referred to as reference.
Cannot Save Recovery Information for Bitlocker in Windows 7
http://blogs.technet.com/b/askcore/archive/2010/02/16/cannot-save-recovery-information-for-bitlocker-in-windows-7.aspx
After configuring the settings, we can use security filtering or WMI filtering to apply the policy to specific computers.
Regarding this point, the following blog can be referred to for more information.
Security Filtering, WMI Filtering, and Item-level Targeting in Group Policy Preferences
http://blogs.technet.com/b/grouppolicy/archive/2009/07/30/security-filtering-wmi-filtering-and-item-level-targeting-in-group-policy-preferences.aspx
Best regards,
Frank Shen

Server Group not working when one of Job Servers is down

I have a Server Group of two job servers. They have the same version 12.2.2.3 and are attached to the same Local Repository of version 12.2.2.0.
When I execute a batch job (from the Designer) and explicitly specify on which job server it should run, it works fine for either job server. Also, when I specify Batch Job execution on Server Group, it works fine.
However, when I shutdown one of the Job Servers, and then try to execute the job on Server Group, I'm getting two error messages, BODI-1111011 that one Job Server is down, and BODI-1111009 that another Job Server has failed.
At the same time, when I check the Server Group in the Management Console, it shows that the allegedly failed Job Server is in the status green.
That error is not reflected in a job server eveng log, nor there is anything written to webadmin log, not in the job trace (the latter isn't created at all).
Is there anything I can do at this point except raise a support message?

The issue was with different users for Local Repository in Admin and Job Server config. I discovered it when trying to run the job from Admin Console. Designer is probably not the best diagnostic tool for such kind of issues.

How to send the data to SQL server group in Reciever JDBC scenario

Hi Experts,
I have done IDOC to JDBC scenario, the data is successfully going to SQL server in reciever side, but in reciever side they got local server and other server groups also there, the data ia m sending is going to local server group, so where can i mention the other server group name in reciever JDBC adapter.
Kind Regards,
Praveen.

Hi,
Except if there is a way to specity this in the URL field (either using a suitable properties or a specific syntax), I do not see such feature in the JDBC adapter ... I'm not familiar with SQLServer group concept, but maybe you could use multiple receivers instead ?
So far, I haven't been able to see how to use groups in the URL ... Maybe there is way to trigger replication at SQLServer level ? Stored proc ?
Chris
Edited by: Christophe PFERTZEL on Dec 9, 2009 11:26 AM

Microsoft forefront protection 2010 for exchange server is unable to perform the requested function

I get the following error any time I try and change any setting:
Microsoft Forefront Protection 2010 for Exchange Server is unable to perform the requested function. This may be becuase Microsoft ForeFront services are unavailable. Ensure that all Microsoft ForeFront services are running and that Windows Powershell is
functional.
I have installed the latest rollup for SEP. Rollup 5 I think. I have rebooted the server. I have checked the permissions and they are correct.
Any thoughts?

I posted a PowerShell-based workaround in the other thread:
https://social.technet.microsoft.com/Forums/forefront/en-US/1ccb9a5e-4b08-4f6b-a4bd-32cf5f2cd2b0/cannot-save-any-changes-microsoft-forefront-2010-for-exchange-is-unable-to-perform-the-requested?forum=FSENext
Mike Crowley | MVP
My Blog --
Baseline Technologies

What most affects Flash performance?

What components, hardware and software, affect Flash
performance the most?
I believe that video cards have no bearing, but have
recently come into a debate over the RAM vs processor.
Also, on Windows XP, any tricks to bring up performance
speed?
Thanks in advance,
michael

Do you mean in reference to authoring or the player?
Dan Smith > adobe community expert
Flash Helps >
http://www.smithmediafusion.com/blog/?cat=11
http://www.dsmith.tv
"msimmons55" <[email protected]> wrote in
message
news:fs8fd5$nvi$[email protected]..
> What components, hardware and software, affect Flash
performance the most?
>
> I believe that video cards have no bearing, but have
recently come into a
> debate over the RAM vs processor.
>
> Also, on Windows XP, any tricks to bring up performance
speed?
>
> Thanks in advance,
> michael
>

RFC Server Groups affecting RFC performance

Similar Messages

Maybe you are looking for