October 19, 2013

SAP NetWeaver 7.0 Dual Stack Refresh Procedure -6

SAP NetWeaver 7.0 Dual Stack Refresh Procedure -Part 6

8.3 Post Refresh Steps

For QBI and SB1, the application server must be reinstalled. For SAPINST to work properly during the installation, we must request that the LDAP integration be disabled. Raise a ticket with GLMSE early to ask for this, or it will create a delay near the end of the refresh.

8.3.1  ABAP Stack  

•  Run “stopsap”

•  Upgrade the kernel to the proper level. This will avoid the adjust Nametab error that indicates SAP is not  communicating with the database. Depending on the system status, JSPM may be used to update the kernel to the correct level.

I. The current kernel for BI is patch 39 for 701. It can be found in /bswork/exe_AIX_701_39_nuc and     /bswork/exe_AIX_701_39_uc. There are also Linux versions available if the system being refreshed has Linux application servers

II. Copy the NUC version to /sapmnt/<SID>/exe 
               cp –pR /bswork/exe_AIX_701_39_nuc/* /sapmnt/<SID>/exe/

III. Copy the NUC version to /usr/sap/<SID>/DVEBMGS<INSTNR>/exe 
               cp –pR /bswork/exe_AIX_701_39_nuc/* /usr/sap/<SID>/DVEBMGS<INSTNR>/exe/
IV. Double check that the UC version is available under /sapmnt/<SID>/exe/uc/rs6000_64 for AIX and /sapmnt/<SID>/exe/uc/linuxx86_64 for Linux application servers. 
               Copy as necessary from the UC versions under /bswork.

• Start the database only by running the startdb command as <sid>adm.

• Run refresh perl script with the “clean” option to clean up tables and change what can be done from a script/DB2 command line perspective. All actions at this step are controlled through the clean.ini configuration file.

The message
“SQL0100W  No row was found for FETCH, UPDATE or DELETE; or the result of a query is an empty table.  SQLSTATE=02000” can be ignored. This indicates the table had no affected records when executing the SQL delete statement. When using the scj_refresh.pl script, this is returned with code 256.

 Double check any errors with return code 1024.
I. BI System
scj_refresh.pl clean <SID>

•  Run post steps script to repopulate, optionally, the user records, RFCs, PSE, ports, and partner profiles
   This may introduce issues with the Java stack as the ABAP user records and the Java UME will be out of  sync.
I. BI System
scj_refresh.pl restore <SID>

•  Enable batch processing
Double check the profile overall – ensure the correct instance and DEFAULT profiles are in place.
Restore the batch processes number from prior to the refresh
Also check the profile for the SSO parameters
login/accept_sso2_ticket = 1
login/create_sso2_ticket = 2

•  Startsap

•  Apply license – Not always  required.
   SAPINST attempts to install a temporary license, however this may fail
   Apply license as required depending on license type (old vs. new)

•  Old license application procedure
  saplicense at the command line

•  New license application procedure

 There is a new command line tool available in recent patch levelsUse “saplikey” at the command line to   import new style keys instead of needing to unlock SAP* and   follow this process of breaking into the system. This process needs to be verified.

•  Log into ABAP and add system message (SM02)

•  Update transaction SCC4
 Changes without automatic recording
 Changes to Repository and cross-client Customizing allowed
 Protection level 0: No restriction

•  Check SM28 for installation problems

•  Run RSBTCDEL to delete canceled and finished batch jobs

•  Check SM12 for locks – delete any obsolete locks
•  Check SM13 for updates – delete any obsolete updates
•  Check Spool/TemSe
  o  SPAD --> Admin --> Consistency check of spool database
  o  Delete all button
  o  SP12 --> TemSe Database --> Consistency check
  o  Delete all button

•  BDLS conversion – Execute in the foreground on a windows server (i.e. USRACIB05). There have been  issues running some of the BDLS conversions in the background with Business Warehouse systems. Typically, there are workbench request prompts that cause the program to “hang” if run in background.

 I. BI system

SB1 Refresh:

QBI Refresh:
PBICLNT100 > QBICLNT100 (approx. 1 hr)
QA2CLNT100 > QA2CLNTOLD (approx. 1 hr)
PRDCLNT100 > QA2CLNT100 (approx. 30 hrs)
PA1CLNT100  > QA3CLNT100 (approx. 1 hr)
PE2CLNT100 > SE2CLNT100 (approx. 1 hr)
PP1CLNT100 > SP1CLNT100 (approx. 1 hr)
PY1CLNT100 > SY1CLNT100 (approx. )

XB1 Refresh:

SBI Refresh:
DB1CLNT100 > SBICLNT100 (approx 10 min)
DR1CLNT200 > XR1CLNT200 (approx 10 min)
XA1CLNT100 > XA1CLNTOLD (approx 90 sec)
DA1CLNT100 > XA1CLNT100 (approx 90 sec)
DP1CLNT200 > XP1CLNT200 (approx 5 min)
DE2CLNT100 > BE2CLNT100 (approx 5 min)
DY1CLNT200 > BY1CLNT200 (approx  min)

•  This is now included in the scripts. Modify table RSLOGSYSDEST and update the RFC destination entries to correspond with the updated logical systems from the BDLS conversion step.

•  This is now included in the scripts. Update table RSADMINA with correct user for BW_USER field.
 o  Ensure ALE<SID><CLNT> user is used and password is known and updated in the appropriate places.

•  Check Partner Profiles in WE20 transaction
 o  Check Partner Status to ensure the profiles are active. We have seen that under certain circumstances,  this is switched to “I” or inactive. Replace with “A” for active and save.

•  RSA1 for BI system
 o  Run transaction RSA1
    Double click Source Systems
    Right click each logical system
    Select Restore from menu to re-connect
 o  Log into SCM system
   Run transaction RSA1
   Double click Source Systems
   Right click on BI logical system
   Select Restore from menu to re-connect
 o  Notes: Both systems need to be open during the connection restore processes. If the restore process  fails, it may be due to any of the following.
>  RFC destination definition issues
>  ALE user permission issues
>  ALE user password issues
>  WE20 partner profile settings incorrect
>  WE21 TRFC port settings incorrect
>  One or both systems locked via SCC4 settings
>  ALE user needing to be Dialog to allow restoration
>  Others I can’t remember right now

•  The BI team needs to ensure the assignment of Source System IDs is correct in order for the RSA1 connections to be re-established.
 o  Navigate to RSA1
 o  From the Tools menu, select “Asgnmt of Srce Syst. To Srce Syst. ID” The following screen is displayed,  however it is not Basis responsibility to fix any entries here.

•  Update PSE
 o  This should be covered by the scripting at this point.
 o  Check STRUSTSSO2 for the appropriate system certificates
 o  Import PSE data that was exported during the preparation phase. This will load the appropriate  certificates again.
 o  If the “Own Certificate” is incorrect, run SSFPSEMAINT through SE38 to generate the PSE

  o  Add any missing certificates to the Certificate List and appropriate entries to the Access Control List  (ACL)
•  Update “Connected Portals”
 o  T-code SM30
 o  Table “RSPOR_T_PORTAL” and click “Maintain”

 o  Update name of system and URL values appropriately.

•  Maintain the SLD connection
  o  From transaction SLDAPICUST, correct the username and password as needed.

•  Ensure the Maintenance Package is scheduled
         Execute SDCCN and delete the maintenance package and refresh package that could exist there.
         Create new by Selecting “Task --> Create --> Maintenance Package”

 Repeat for “Refresh Service Definitions.”

        Note: “Refresh Sessions” should not be running, delete it if it exists.
•  Update TMS landscape appropriately (Client 000)
 o  Integrate into appropriate domain (QA systems).
 o  Check SECSTORE transaction for stored ‘bad’ passwords for TSMADM, etc.
 o  Run RDDNEWPP in working client to schedule the transport jobs


>  Connect to the various precalculation servers and reset the status. The existing precalculation server  definitions should be removed and the connections added once more. Check the RFC destinations also since if the RFC already exists, you cannot add the server in RSPRECADMIN. This is based on the name given during the server addition.
>  Production server USRACIA73

>  QA server USRACIQA000
>  Dev server USWAXDDA004
>  All three servers are configured in PBI. The two test servers are configured in QBI and SB1. Only the development server is configured in DB1. The sandbox systems only use precalculation on a need basis, so maintenance there is not required unless specifically requested from the BW team.

Will update soon... Check next post...

1 comment: