While starting the SWPM tool for a system installation I came across this message:
./sapinst SAPINST_USE_HOSTNAME=hostname
[==============================] - extracting... done!
SAPinst build information:
--------------------------
Version: 749.0.23
Build: 1763064
Compile time: May 19 2017 - 07:46:46
Make type: optU
Codeline: 749_REL
Platform: linuxx86_64
Kernel build: 749, patch 217, changelist 1761678
SAP JRE build: SAP Java Server VM (build 8.1.030 25.51-b02, Apr 26 2017 16:01:22 - 81_REL - optU - linux amd64 - 6 - bas2:288514 (mixed mode))
SAP JCo build: 3.0.16
SL-UI version: 2.5.30
SAP UI5 version: 1.42.4
Exe directory: /share/packages/SWPM/tmp/sapinst_exe.29580.1497432865
load resource pool /share/packages/SWPM/resourcepool.xml
INFO 2017-06-14 09:35:17.155 (root/sapinst) (SLPCommunicator) [/bas/749_REL/bc_749_REL/src/ins/SAPINST/impl/src/guiengine/SLPMonitoringStatemachine.cpp:1355]
********************************************************************************
Access URL for GUI is
https://hostname:4237/w8RbRSZrzplmfEY9/docs/index.html
Logon users: [root]
********************************************************************************
sapinst was hanging here without starting the GUI window
It looks like starting with SWPM PL 19, the installer can start now in two ways:
– new web browser based, the link where this can be accessed appears in the sapinst log
– or the classic Java GUI based. This is not available by default anymore, in order to access this in the old way, you have to use sapinst with the command line option SAPINST_SLP_MODE=false
Recently, searching the developer traces I found these messages, right at the beginning of the log:
RSecSSFs: Configuration value "rsec/ssfs_datapath" from profile explicitly overwritten by different value in environment variable RSEC_SSFS_DATAPATH [rsecssfs.c 4677]
RSecSSFs: Configuration value "rsec/ssfs_keypath" from profile explicitly overwritten by different value in environment variable RSEC_SSFS_KEYPATH [rsecssfs.c 4686]
RSecSSFs: Reverting overwriting of profile by different environment (disabled in dw) [rsecssfs.c 4696]
I checked the SSFS configuration, more precisely the instance parameters:
rsec/ssfs_keypath = $(DIR_GLOBAL)$(DIR_SEP)security$(DIR_SEP)rsecssfs$(DIR_SEP)key = /usr/sap/SID/SYS/global/security/rsecssfs/key
rsec/ssfs_datapath = $(DIR_GLOBAL)$(DIR_SEP)security$(DIR_SEP)rsecssfs$(DIR_SEP)data = /usr/sap/SID/SYS/global/security/rsecssfs/key
And also the environment variables on the central instance:
Even if the two paths /usr/sap/SID/SYS/global/security/rsecssfs and /sapmnt/SID/global/security/rsecssfs point in the same location, it looks like the systems sees them differently that is why in also the environment variables on the application servers need to be changes to the same name path name /usr/sap/SID/SYS/global/security/rsecssfs/key and /usr/sap/SID/SYS/global/security/rsecssfs/data.
The situation of the file system full for /usr/sap is frequently met. Probably that the first thing that any admin would do is to check if he can cleanup any core files on the FS.
In order to have an overview and also to avoid this manual check/cleanup, I created a script to check the existence, time stamp and size of such file and send this information via email.
Also if the file is older than seven days this will be deleted, having enough time to analyze the cause for the generation of the file.
Here is how the script looks like:
#!/usr/bin/ksh
cd /usr/sap/SID/DVEBMGS*/work
FILE1=$(ls -l core |grep -v grep | grep -cw core)
FILE2=$(find . -name core -mtime +7 | grep -v grep | grep -cw core)
if [ $FILE1 != 0 ]
then
echo "There is : $(ls -l | grep -cw core) core
with the name : $(ls -l core)
and size : $(du -h core)
space left on FS :
$(df -h core)
$(hostname), $(date)" |
/usr/lib/sendmail -v email@domain.com
else echo "no core - do nothing"
fi
if [ $FILE2 != 0 ]
then
(rm core) && (echo "A core file older than one week has been deleted on $(hostname), $(date)" : [$(ls -l core)]) |
/usr/lib/sendmail -v email@domain.com
else echo "no core older than one week"
fi
This is how such an email should look like, if any cores are found:
There is : 1 core
with the name : -rw------- 1 sidadm sapsys 90001069 Mar 19 16:12 core
and size : 86M core
space left on FS :
Filesystem size used avail capacity Mounted on
hostname:/usr/sap 1.0G 175M 849M 18% /home/sidadm
hostname, Thu Mar 24 10:10:00 CET 2017
Introscope EM is a third party standalone engine that is installed under Solution Manager and used to monitor the performance of Java applications.
In this article I will present its installation under Solution Manger 7.2 on Linux platform. The version 10.1 is anyway only compatible with Solution Manager 7.2 (>=SP03). But the minimum version that can be installed on Solution Manger 7.2 is 9.7 SP01.
First of all, three packages are needed for the installation:
Introscope Enterprise manager
Location for download: https://support.sap.com/swdc -> Software Downloads -> Support Packages and Patches (2nd tab) -> By Category –> SAP Technology Components -> SAP SOLUTION MANAGER -> SAP SOLUTION MANAGER 7.2 -> WILY INTROSCOPE -> WILY INTROSCOPE ENTPR MGR 10.1 -> (your OS for the enterprise mgr)
2. Introscope Management Modules to be installed on top of Introscope Enterprise Manager
Location for download: https://support.sap.com/swdc -> Software Downloads -> Support Packages and Patches (2nd tab) -> By Category –> SAP Technology Components -> SAP SOLUTION MANAGER -> SAP SOLUTION MANAGER 7.2 -> -> WILY INTROSCOPE -> WILY INTRO EM MGMT MODULES 10.1 -> OS independent
3. Third-party external component package files based on the release
Now that we have the packages, we can create a directory where to place and unzip the Introscope EM. Here we will also put the external component files, as the installer will automatically identify these during the installation.
Then start the installation in silent mode with sidadm user (sidadm is the admin user of the Solution Manger instance):
./introscope10.1.0.15linuxAMD64SAP.bin -Djava.awt.headless=true
Preparing to install...
Extracting the JRE from the installer archive...
Unpacking the JRE...
Extracting the installation resources from the installer archive...
Configuring the installer for this system's environment...
Launching installer...
Silent installer completed with status of: SUCCESS. For more details, consult install logs at: /usr/sap/ccms/apmintroscope/install. Please wait for the command prompt to return...
And the installation is completed. For me the installation did not take more than 5 minutes.
You can check the installation logs under /usr/sap/ccms/apmintroscope/install
In case the installation wasn’t successful you can find more hints in the logfile: /usr/sap/SID/willy_ent_mng/silent.install.failed.txt
The default directory for Introscope EM is /usr/sap/ccms/apmintroscope. Here you can also find the script for start/stop/check state.
After this is installed, install the module. This is done by copying the module zip in the directory of wily: /usr/sap/ccms/apmintroscope and unzip WILYISMM00P_1-70001722.zip here directly.
Now let’s start it:
cd /usr/sap/ccms/apmintroscope
./EMCtrl.sh status
./EMCtrl.sh status: Enterprise Manager stopped
./EMCtrl.sh start
./EMCtrl.sh start: Starting Enterprise Manager...
(Please wait a few minutes for initialization to complete.)
Please check log file for more details.
The link should be reachable: http://hostname:8081/webview
There are two users available for login. If you did not change their credentials before installation the default passwords are:
user Admin, passwd: Admin89
user Guest, passwd: guest12
You can find other related EM articles on this site at:
