Sunday, 17 June 2012

SAP BASIS System Monitoring

1) OS level monitoring
 a) file system monitoring
 b) Network connections: SAP Router
2) SAP System monitoring

 a) Check DB and SAP services
 b) Working with tcodes to monitor SAP system
-*-
1) OS level monitoring
a) File system monitoring
Perform quick check on the growth of oraarch (% free) and also look for mapped / partition size(%free) if the threshold goes to  high take necessary actions Ex, Backup / Delete.
b) Network connections: SAP Router
Check your SAP Router connection by login to the system and error logs present at
-*-
2) SAP System monitoring
a) Check DB and SAP services
Win:
open services.msc => look for SAP and oracle services are in "start" status
Unix:
ps -ef | grep ora
ps -ef | grep db2sysc
                          -*-
Oracle:
login with sidadm
sqlplus /nolog
connect / as sysdba
If connection done without any error then it's good move forward.
In case of any issue open below file contains information about startup, shutdown, error, etc
Alert_SID.log -> Location : /oracle/SID/saptrace/background
DB2:
login with sidadm
db2 list applications
For logs and alerts execute : db2diag -f
-*-
Working with tcodes to monitor SAP system:
SM51 List of SAP Systems
- Display list of active/up/running application servers that have registered
- Also you can also manage, and display the status of the users (SM04), Gateway (SMGW), ICM (SMICM), SNC, Release notes, Remote logon, System log entry (SM21) and browse SAP directories from "Goto" option
- Also you can check what WP's are defined for every listed Instances from the "message type" field  ex, DIA,BG,UP,SPOO etc., Note:: From release 7.1 no "Enqueue WP" will not display to monitor use ensmon tool.
- If you are logged on DI1 and want to check something on DI2 then select the system DI2 from SM51 - double click and enter the tcode
SM50 Work Process Overview
Monitor this section and look that enough WPs are in waiting status to hanlde next user request
Display list of WPs (DVEBMGS) - look for Status and Action tab
You can read dev_wp* file directly from here without going to OS which is present at /usr/sap/SID/DVEBMGS00/work/*
To display the file clieck on "Display file" Icon or ctrl+shift+f8 will open the file
Also we can cancel any active process from Process menu with core generate file or wihout core will not generate log file
SM66 Systemwide Work Process Overview
Global workprocess is very usefull when we have more than 1 DI this will show all WP (CI,DI) WP status
look for all WP processes and observe error (red) or in stale state,
We can also display only stop process goto => process list => select process => check status
Restart the stop process by going to stop wp app server (sm50) => select wp => process => Restart after error => yes
SM12 Display and Delete Locks
At selection only give client no and execute look for old entries sorting by time, It may possible to observe 6-7 days old logs or more but before taking any action look in details for user status and get details from AL08/SM04 for current action, more you can check if any Background job is running with that user name from (SM37) check if any jobs are active for that user if nothing is happening check for any stale connection from SM04/AL08 and disconnect the user from SM04 then check if the lock is clean if not then delete this manually
To get details select lock and click on "Details" (F2) screen will popup look for:
User,Sysno,WP,tcode to analyse further by check wp trace file and system log (SM21)
Get details on lockmode (click F1) and Cumulative Counter
SM13 Administrate Update Records
Check the update system which has to be "Update is active" select status all and execute
Select if any "Error" status click on "Update Header" (hat icon) look for Error details,User,Tcode,Update server to analyse further also double click on error details it will give more details Ex, ABAP dump etc.,
AL08 Users Logged On
Check no of logged users for all instances (CI,DI)
SM04 User List / AL08 - Global user over for multiple Instance
Find stale user connection before taking any action check if there's no job,program or nothing is running by selecting the user - select sessions
Sometime you w'll see users without any sessions in this case look for "Type" this will be a : RFC, Plugin Http etc.,
SM37 Overview of job selection
Analyse: cancelled jobs
Select canceled job and clieck display job log => The log contains every step details which is completed and cacelled look for the one which is canceled => look if this is due to some functional issue, ABAP Dumps, program dependent, variants etc.,
To get more details double click on the job => go to job details => look for target server and wp to analyse further or select job go to edit => choose details => to view client,system,timings etc.,
To search jobs by program name Ex- RSPO1043 to find the job name
We can also view the spools generated by job, modify steps, program, variants, schedule timings etc.,
To analyse the long running jobs select the job => goto => job statistics => analyse by - res time, db req time, wait time, transfer data double click to analyse more
Also look for standard SAP jobs (start with SAP*) were running sucessfully if not then need to do start or if failed then analyse : SAP_REORG_JOBS, SAP_REORG_SPOOL, SAP_REORG_BATCHINPUT, SAP_REORG_ABAPDUMPS, SAP_COLLECTOR_FOR_JOBSTATISTIC,
SAP_COLLECTOR_FOR_PERFMONITOR etc.,
ST02 Setups/Tune Buffers
World of SAP buffers...
Look closely for every parameter specially those which is displayed in red so wherever the red alert is there change the parameter value depends on requirement but before that analyse and display the current value by double click => current parameters => to see all parameter for all area select all or on main screen current parameters
Keep in mind that each buffer holds different values
If you want to get your hands more sirty on this then go to => detail analysis menu => look every values under SAP runtime buffers and to see DB hot ration and quality for the week goto main screen => summary of all buffers (shift+f2) scroll down untill you see "Development of SAP buffers"
Use below command for swap space requirement for SAP application server:
sappfpar check pf=/usr/sap/<SID>/SYS/profile/<Inst_Prof> nr=<Sys No> name=<Systemname> | more
look at startup date if you want to know from how long your SAP system is up
ST06 Operating System Monitor
look for user utilization, Disk res. time, swap info etc., double click on utilization user to get hourly details on CPU utilization
Go to "Detail Analysis" to get HW Info, OS logs, Memory, Filesystem, LAN etc.,
If this reports to high CPU utilization check for the highest utilization process in Unix / Linux by "TOP" select the PID and do the analysis from "SM50" also check SM66 / SM66 CPU filed by clicking on white watch icon and find the highest utilizing CPU, check under SM37 by secting only active process sometime may be of jobs execution also bring high utilization
ST04 DB Performance Monitor
Take a look at Buffer Quality and Hit Ratio a healthy system always point to 98.xx
Analysing poorest SQL queries goto => sql cache => delete filters => take a look of total execution column
monitor DB growth, Tablespaces,Missing Index etc.,
ST22 ABAP dump analysis
The theory of analysing the dumps are not depend only with ST22 sometime you may also need to work with => SM21,SM50,SM37,STAD,ST03N,ST02,RZ10 etc, and finally searching in SAPNotes for proper understanding to solve it.
Quick Look-up Fields => User and T-code, What happened, Error Analysis etc.,
ST03 Workload and Performance Statistics
Monitor Response time for Transacton : Select app server under expert mode => select duration, date-time, day etc., => Select transaction profile => standard => check for poorest t.response time filed and observe if this t-code is coming every time in poor response
Response time from user => Expert mode => user and settlement => user profile => look for t.response time from the filed
Also use STAD to collect more data by user, tcode, program with with multiple instance selection and more
DB02 Tables and Indexes Monitor also called as Static Database Monitor (DB02_MSS - MS.SQL)
Monitor DB size, Tablespaces size, Tables and Indexes (also find missing indexes here)
List of default tablespaces:PSAPTEMP, PSAPUNDO, PSAP[SID], PSAP[SID]620, PSAP[SID]UDR and SYSTEM
A Refresh button is provided to update the information.
DB analysis: Offers a number of options to execute checks on the database
Detail Analysis : Executes a consistency check for a specified table and returns detailed information about the table.

Analyse Missing Indexes:
Choose the Database analysis button => A list of available database checks is displayed => To start the search for missing indexes, choose DB <-> ABAP/4 Dict. In the dialog box that appears choose Recheck => When the check has finished, the results are displayed on the screen.
DB12 DBA Backup Logs
Monitor: DB and Archive Backup logs
Check for logs on failed backup, duration of complete backup, size of last backup Database, backup device type information etc.,
DB14 Display DBA Operation Logs (Oracle)
Logs generated during execution of BRSPACE, BRCONNECT, BRARCHIVE, or BRBACKUP
SARFC Server Resources for Asynchr. RFC
Monitor available Server Resources for Asynchr. RFC in R3 system (Ex-R3 connected with BW)
SMQS Registration of Destinations
qRFC Monitor - check scheduler information and status
SMQR Registration of Inbound Queues
SM59 RFC Destinations (Display/Maintain)
Create, Change, Delete, Test RFC's using ABAP connection R3 to R3 or R3 to BW, APO, GTS etc, use TCP to use ex
Check RFC connections perform connection test and authorization test
To check RFC is working Select RFC => Utilities Test => Connection Test or Authorization Test
If it gives connection error check for IP/hostname of target server, system no for Authorization Test check username and password is maintained in remote systemhttp://help.sap.com/erp2005_ehp_04/helpdata/EN/22/0424ba488911d189490000e829fbbd/frameset.htm
SM21 Online System Log Analysis
System logs analysis for local or all instances analyse logs by user, tcode, date and time, WP select problems only and execute all red text display are problems like Dumps, connection errors, Os, DB etc.,look specially for resource related entries
To get logs from all instance goto => system log => choose => select all remote system logs => execute
SM28 Installation Check
Perform this check after restart, kernel, patch, notes, add-ons upgrade or any other critical activity
SMLG
Create, change, delete logon groups
Check for all define logon groups are in green status
To create goto create assignment => logon group name ex: INDIA, provide Instance name =>
Assignment Specific  : properties specific to this 'group-server' pair.
IP address  : Frontend IP connection
Group Specific   : properties specific to this logon group.
External RFC enabled : If external RFC logon is allowed or not.
Application Server Specific - properties specific to this application server.
Response Time  : Threshold of the average dialog response time
User   : Threshold of the number of logon users
SMICM ICM Monitor
Check status of ICM and Active Services: goto => services => HTTP / SMTP with their port nos pointing to right app server
Check for parameter and trace file for error log
For dual stack manage the J2EE Engine from here goto administration => J2EE Instance => Sending a Soft / Hard Shutdown, end process, restart
SOST - SAPconnect Send Requests
View the logs sent by SMTP,FAX etc.,
Look for error details on "No Delivery" and current waiting status
Also execute or repeat send for all and selected requests 

2 comments: