Professional Documents
Culture Documents
ERP Common Errors Solutions
ERP Common Errors Solutions
This documents all UNIX, Oracle, SAP related errors that are worth sharing across teams.
TABLE OF CONTENTS
UNIX 2
ORACLE 4
SAP GENERAL 10
SAP PRESENTATION 16
SAP BACKUPS 20
SAP SECURITY 25
UNIX
4. What is the IP of the Unix server? a) nslookup <SID> : Will use DNS
b) /etc/rc.config.d/netconf
c) /etc/hosts
d) lanscan (get lan0); ifconfig lan0
5. What do the svr* jobs scheduled in EMEA a) svxrj001 job performs a database system check. Check if
systems do? database is up and running.
b) svxrj010 job performs a sapdba -next. This checks if
there are no problems with the max.extents.
6. Activate PRS_BIG and deactivate the old > alter rollback segment prs_1 offline;
rollback segments ...
> alter rollback segment prs_10 offline;
> alter rollback segment prs_big online;
TEMPORARY FIX
If tablespace is urgently getting filled up, can increase
size of an oracle datafile on line without rebouncing.
Alter database datafile
<datafilename> resize <new size>
If forced to kill job in unix, PID in SM66 is the app server PID
(client).
>saplicense -install Check that the license used has expiry date
to install new license. Must delete old license first before 31/12/9999
installing new one.
>saplicense -temp
to install a temp license
11. Changes / Transports to the database server Modify default.pfl parameter to contain:
are not seen from the application server
rdisp/bufrefmode=sendon,exeauto
12. /usr/sap/trans is full Run the scripts :
1) Go to /usr/sap/trans/data
2) /usr/sap/trans/scripts/compandmove <# of days>
/usr/sap/trans/archive/datafiles
SAP GUI
2. Spool table TSP01 reaches limit of 32000 Run RSPO0041 job with transaction SE38 with Basis to check batch job logs
following variant, client : * and check the delete
all print job with min age option and Commit
parameter: 1000 can check the table size with
the following SQL statement in svrmgrl,
Select count(*) from sapr3.tsp01;
3. Print jobs cannot be printed out. Printer configured incorrectly at SAP or Unix Follow proper procedures for printer
level or both. Reconfigure printer. configuration.
4. Print job at SAP & Unix level is completed but Netware or local site infrastructure down.
printer does not print it out. Cannot print from Forward issue to local site to handle.
other windows applications like Microsoft Word
and Excel.
Delete /usr/sap/L71/DVEBGMS00/data/SP*
Clean up stats file via st03
Clean up work process trace files via sm50
Recycle spooler
9. EMEA Printing Support EMEA Printing Servers: etsp0033; etsp3005
remsh etsp0033 -l prtmgr "lpstat -p
ETC920R"
smitty
Note: In order to get a better idea of the reason for the failure the following 3 steps are highly recommended
1. Look into the logs and search for the instance of the first occurrence of any error message instead of looking at the tail of the logs. Very little information is
conveyed and hardly anything can be deduced by looking at the brbackup/brarchieve return code mentioned at the end of the logs. Most of the following
data is catering to the first error message that comes up in the logs.
2. Note that it is the first error msg that really counts and all the errors that crop up after that are usually a follow up of the first error.
3. Note that the /var/opt/ctma/ctm/sysout logs are lot more accurate and better furnished in detail as compared to the logs we see at SAP GUI or
even in the /saparch or /sapbackup directories. So if you are not able to see the error in the usual places look at the sysout logs. [Just go to the dir and
do $ ll | grep [jobname that failed] to see the logs.]
#./sapg900_cleanup_pool.sh +3 var/opt/saparch/prod
#./sapg900_cleanup_pool.sh +10 var/opt/saparch/dvlp
sesu - orasid
svrmgrl
connect internal;
select * from V$BACKUP;
alter tablespace <tablespacename> end
backup;
./removelock.sh SID
5. pgbrclean.sh not found The script is missing from the /sapmnt/SID/exe directory
Copy over the script from the /opt/nb/bin directory and in case
it is missing from there, ftp it from another source to /bin and
copy it to /etc. The backup would normally be successful, so
dont rerun the backup, just fire the script as ora<sid>
6. backint execute permission denied Problem with the permissions of backint. Usually the sticky bit
would be missing.
7. Specified schedule does not exist in the The parmecs file would have been refreshed to the incorrect Work with UIT to make sure correct parmecs
specified class parameters file gets refreshed.
SOLUTION:
1) Ticket to UIT
2) Grant correct roles to OPS$ user:
svrmgrl
connect internal
grant sapdba to OPS$<SID>ADM
grant connect to OPS$<SID>ADM
grant resource to OPS$<SID>ADM
# default: $ORACLE_HOME/sapbackup
backup_root_dir =
(?/dbdump0,?/dbdump1,?/dbdump2,?/dbdump3,?/dbdump4,
?/dbdump5,?/dbdump6,?/dbdump7,?/dbdump8,?/dbdump9,?
/dbdump10,?/sapdata92,?/sapdata94,?/sapdata95,?/sapdat
a96)
12. /var/opt/saparch in pooldisk is 100% full Attempt to delete using
/var/opt/sapg/scripts/sapg900_cleanup_pool.sh + x
/var/opt/saparch
/var/opt/ubkp/scripts/ubkp049_varopt_genfile.sh
generates the listings of files to be backed up in
/var/opt directory
/var/opt/ubkp/scripts/ubkp050_varopt.sh <CLASS>
<SCHED> <HOST> <FILE> backs up the files
/var/opt/ubkp/scripts/ubkp050_varopt.sh sihp8040
d_2w_e sihp8040 ubkp050_varopt
230 the specified class does not exist in the configuration database 1/setup problem, send to UIT
4. Backup failing due to SQL error -942 SAPDBA role has not been defined yet and <sid>adm is not Whenever chdbpass is executed and
granted SAPDBA role. OPS$<sid>adm is recreated, you need to
run sapdba_role script afterwards.
DIAGNOSIS:
1) sesu - ora<sid> 1) sesu - ora<sid>
2) svrmgrl 2) cdexe
3) connect internal 3) sqlplus internal @sapdba_role <SID>
4) select * from dba_role_privs; UNIX
5) OPS$<sid>adm should have CONNECT, RESOURCE,
SAPDBA role assigned to him. If SAPDBA role is not
assigned to him, do the ff.
SOLUTION:
1) sesu - ora<sid>
2) cdexe
3) sqlplus internal @sapdba_role <SID> UNIX
5. <sid>adm can still do svrmgrl and connect to <sid>adm should not be able to connect to the database.
the database Only <ora>sid should be part of DBA group
DIAGNOSIS:
1) more /etc/group | grep dba
2) If you see users other than <ora>sid, request UIT to take
them out of DBA group
bdhm:/var/opt/sapsoe_arch/depot MONITOR
7. zpas doesnt work for some systems Your user account has not been created in the system or your
zpas reports error in the status account on the said system is locked.. Zpas (password
Tried to login using new password that propagation) only works for those systems where your
was propagated by zpas but cannot login account exists. Open ITSM ticket SSC 985.
8. zpas works to propagate new password to a Password you provided zpas is not unique. Password you Provide a unique password.
system but when you login, system asks you provided was one of the last 5 passwords youve already
to immediately change your password. Effect used.
is your passwords across all systems are not
synchronized anymore
9. How to know which policy model a particular AIM is using 2 policy models for hp:
server is subscribed to? 1 - hp_sap_sec@bdhp4176 : This is the highly secured
policy model for all Production systems
2 - hp_sap_std@bdhp4176 : This is the standard policy
model for all non-production systems
To know which :
1) Go to: http://golx4001.na.pg.com/~markj/pmdbs/
2) grep pmd /usr/seos/seos.ini' will return several lines.
Look for the line that says, "parent_pmd = ..." This
will identify the policy model.