Hi Rafal,
Please find below details of the link:-
Symptom
No connection to the message server appears to be available or possible or you want to check or display details of the the message server using the active instances and available LOGON groups directly on operating system level.
In the SAP system, you can do this using the transactions SM51, SMMS or SMLG.
You are looking for a tool to perform these checks directly on operating system level.
Other Terms
Testtool
lgtst
Message server
Logon groups
Load distribution
SM51
SMLG
RZLLICLASS
Weighted round robin procedure
Reason and Prerequisites
The message server runs exactly once for each SAP system.
Its task is to notify all servers belonging to a SAP system (instances) of each other's existence and it can also be contacted by other clients (such as SAPlogon, RFC clients with load balancing) for load balancing or information retrieval.
A prerequisite is that the message server must be uniquely addressed by specifying the message server host name and the message server port.
In the case of the message server port, it is particularly important that the port number assigned to the port name (sapms<SID>, SID=SAP-System-ID) is identical on all participating hosts (for example, 3600/tcp, see /etc/services) for example.
On the SAP System, these specifications are defined in the DEFAULT.PFL global profile:
- rdisp/mshost = <hostname of the message server>
- rdisp/msserv = sapms<SID> (optional specification, naming convention)
When you start a SAP instance, the dispatcher process tries to establish a connection to the message server (with these specifications), to declare the services it provides (DIA, BTC, SPO, UPD, ICM etc.). If the connection fails, a corresponding entry is inserted in the syslog.SolutionThe message server /Logongroup test tool "lgtst" is available for every SAP server platform and every SAP front-end platform.
It recognizes the following call parameters:
lgtst -H <host name message server> -S sapms<SID>
To test the message server connection, you should use the above-mentioned SAP configuration parameters.
If this succeeds, you get a split list:
- List of active SAP server instances (with services)
- List of available LOGON groups (with default instances)
For example:
using trcfile: dev_lg
list of reachable application servers
-------------------------------------
[oss001_O01_01] [oss001] [147.204.100.8] [sapdp01] [3201] [DIA VB ENQ BTC SPO ]
[oss006_O01_01] [oss006] [147.204.100.14] [sapdp01] [3201] [DIA ]
[oss005_O01_01] [oss005] [147.204.100.13] [sapdp01] [3201] [DIA ]
[oss003_O01_01] [oss003] [147.204.100.7] [sapdp01] [3201] [DIA ]
[oss004_O01_01] [oss004] [147.204.100.11] [sapdp01] [3201] [DIA ]
[oss000_O01_01] [oss000] [147.204.100.6] [sapdp01] [3201] [DIA ]
[oss002_O01_01] [oss002] [147.204.100.10] [sapdp01] [3201] [DIA ]
list of selectable login-classes with favorites
-----------------------------------------------
[1_PUBLIC] [147.204.100.10] [3201] [640]
[2_JAPANESE] [147.204.100.7] [3201] [640]
[3_RFC] [147.204.100.8] [3201] [640]
Evaluation:
- oss001_O01_01 is the central instance of the system O01 (usually the following message server also runs here ==> rdisp/mshost=oss001, rdisp/msserv=sapmsO01)
- all other instances are pure dialog instances
- for the LOGON group 1_PUBLIC, the instance oss002_O01_01 is currently recommended, similarly oss003_O01_01 for 1_JAPANESE
The data on the instance list (partial list 1) is generated on the basis of the above-mentioned dispatcher message to the message server.
The data on LOGON groups (partial list 2, if available) is calculated by the SAP system for load balancing (provided that this was configured; Transaction SMLG; note 26317).RFC load balancing procedure
As of Release 6.10, new options have been added to the testing of the RFC load balancing. Details on the new RFC load balancing procedure are provided in note 593058.
- -G <name> name of a logon group
- -C loop counter
- -Y delay of n milliseconds between each call
Check the load balancing within the logon group 3_RFC for six calls:
lgtst -H <host> -S <serv> -G 3_RFC -C 6
Group 3_RFC : [147.204.100.8] [3201] [640] [TRUE] [oss001_O01_01]
Group 3_RFC : [147.204.100.14] [3201] [640] [TRUE] [oss006_O01_01]
Group 3_RFC : [147.204.100.8] [3201] [640] [TRUE] [oss001_O01_01]
Group 3_RFC : [147.204.100.14] [3201] [640] [TRUE] [oss006_O01_01]
Group 3_RFC : [147.204.100.8] [3201] [640] [TRUE] [oss001_O01_01]
Group 3_RFC : [147.204.100.14] [3201] [640] [TRUE] [oss006_O01_01]
Here, you can see that the two servers are used alternately for the RFC calls. The field [TRUE] specifies that RFC load balancing is activated for this group; the last field specifies the server to which the next request is assigned.
lgtst -H <host> -S <serv> -G 1_PUBLIC -C 3 -Y 1000
( -Y 1000 , that is, a wait time of one second between the calls)
Logon data for logon classes:
------------------------------
Group 1_PUBLIC : [147.204.100.10] [3201] [640] [FALSE]
Group 1_PUBLIC : [147.204.100.10] [3201] [640] [FALSE]
Group 1_PUBLIC : [147.204.100.10] [3201] [640] [FALSE]
You see here that RFC load balancing is not activated (FALSE). The request is assigned to the relevant 'Favorite Logon Server'.
A new -X option, which returns all available information about an RFC logon group, is supported as of Release 6.20. In this case, the definitions are also displayed in the RZLLICLASS table. You can also see which application servers are defined in this group. Of course, you can also link the -X option with the other options, such as -C and -Y, to test the load balancing.
lgtst name=O01 -X 3_RFC
Logon data for logon classes:
------------------------------
Group 3_RFC : [147.204.100.8] [3201] [640] [TRUE] [oss001_O01_01]
group-id = 2
time to reread = 180 (Seconds)
logon reread = 200 (199) (Number of calls before the next read)
fav type = BEST (defined load balancing procedure)
first read = 1049189654 (Tue Apr 1 11:34:14 2003)
last read = 1049189654 (Tue Apr 1 11:34:14 2003)
next read = 1049189834 (Tue Apr 1 11:37:14 2003)
appl server (name [quality,deltaquality] (Server in the group)
oss001_O01_01 [32,-2]
oss006_O01_01 [81,-2]
During the load balancing procedure ROUND_ROBIN or WEIGHTED_ROUND_ROBIN, the system also issues the internally determined random number for the application servers. This number determines the sequence of the selection of servers if the quality numbers of several servers are identical.
(Release 6.20 as of patch level 40 of the RFC library).
appl server (name [quality,deltaquality,random]
oss006_O06_01 [32,-2,789] => 1.Request
oss006_O01_01 [32,-2,456] => 2.Request
The weighted round robin procedure is only available as of 7.00 and a
certain patch level (also see Note 1112104).
If an alternative IP address was maintained for an application server, then this address is specified behind the name (Release 6.20 as of patch level 1118).
oss006_O06_01 [32,-2] [IP=194.204.100.8]
Documentation
You will find the detailed description in the SAP Online documentation. To do this, go to the help portal (http://help.sap.com) and select NetWeaver -> Release 04.
In the SAP library, select SAP Netweaver -> Solution Lifecycle Management -> System Management -> Tools for System Monitoring -> Monitoring and Administration of the SAP Message Server.
Alternatively, you can enter the following URL directly:
http://help.sap.com/saphelp_nw04/helpdata/en/40/c235c15ab7468bb31599cc759179ef/frameset.htm
Let us know if your still seeing the issue.
regards,
ram