HIGH AVAIALABILITY MAXDB manul FAILOVER test Failed and gives error

[Former Member]

Hello,

we have installed EPR2005 for our production server and we are testing the High Availabilty SUN cluster Environmnet. we are facing problem in MAXDB starting and stopping and all. The manual Stop and Start testing it was working fine earlier days now it is giving problem, I have given the error which i got while testing DB , the actual script will inititate the SAP db start and CI start.

Please some one help me in this reagard.

RR 11779 NISERVER Error during startup:'dlopen failed:ld.so.1: niserver: fatal: /s'

WNG 12457 XSERVER NI Init: Automatic niserver start failed

WNG 12453 NISSLSRV NISSL Init: dlopen failed:ld.so.1: nisslserver: fatal:

12902 XSERVER started, 'X64/SUN 7.6.00 Build 035-123-139-084'

-

---

Wed Jan 2 13:36:31 IST 2008

Starting database

Starting Database Instance...

Error! Connection failed to node pisdbprd for database PIS:

-24700,ERR_DBMSRV_NOSTART: Could not start DBM server.

-24832,ERR_SHMNOTAVAILABLE: Shared memory not available

-24827,ERR_SHMALLOCFAILED: ID /sapdb/data/wrk/PIS.dbm.shi, requested size 2592

Opening Database...

Error! Connection failed to node pisdbprd for database PIS:

-24700,ERR_DBMSRV_NOSTART: Could not start DBM server.

-24832,ERR_SHMNOTAVAILABLE: Shared memory not available

-24827,ERR_SHMALLOCFAILED: ID /sapdb/data/wrk/PIS.dbm.shi, requested size 2592

It is quite urgent, if some one post solution your help would be appreciateable.

Regards,

Sankar.G

BASIS CONSULTANT

[markus_doehr2]

Can you see the given path on that node? Are you on Solaris 9 or 10?

Markus

[Former Member]

Markus,

for the question u have replied is on the Solaris 10 Zones,

we haved checked the path wrk dir , there we have the .shi fil

Please help us in this regarding.

Mahesh

[markus_doehr2]

Did you create a project (/etc/project and /etc/user_attr) file for the users?

--

Markus

[former_member229109]

Hello Sankar.G,

-> Please let me know the structure for the database software directories on node A & node B, see SAP Note No. 803452

Did you have problems with database installation in the cluster Environment?

Did you install the database server software with the <independent_data_path> located on a local file system?

-> Please update with output of the following commands on the node pisdbprd, logon as user in the sdba group:

more /etc/opt/sdb

dbmcli -s inst_enum

dbmcli -s db_enum

sdbregview -l

xinstinfo PIS

dbmcli -d PIS -n pisdbprd -u control, <control>

< enter>

-> Please logon as root user to node pisdbprd & run the following commands < <independent_program_path> = /sapdb/programs by default >:

cd <independent_program_path>/bin

./sdbverify > check_instal.txt

& let me know the results.

-> Please run the following steps & let me know the results < node pisdbprd >:

A) Execute the following program < <dependent_programs>= /sapdb/PIS/db by default >:

<dependent_programs>/pgm/dbmshm UNLOCK /sapdb/data PIS

dbmcli -d PIS -n pisdbprd -u control, <control>

< enter>

If you are not able to connect to the database with the reported errors above after A) =>

B) Execute the following program < <dependent_programs>= /sapdb/PIS/db by default >:

<dependent_programs>/pgm/dbmshm DELETE /sapdb/data PIS

dbmcli -d PIS -n pisdbprd -u control, <control>

< enter>

If you are not able to connect to the database with the reported errors above after B) =>

C) * Rename the files /sapdb/data/wrk/PIS.dbm.shi & /sapdb/data/wrk/PIS.dbm.shm

• Try to connect to the database

dbmcli -d PIS -n pisdbprd -u control, <control>

< enter>

-> "the actual script will inititate the SAP db start and CI start."

Did you write this script or got the script from vendor?

Did you create the OSS ticket, where this issue was reported?

Thank you and best regards, Natalia Khlopina

Edited by: Natalia Khlopina on Jan 3, 2008 8:11 PM

[Former Member]

Hi Natalia

>/pgm/dbmshm DELETE /sapdb/data PIS by executing this the database is up,

Actually u have given this solution for someother also.

Thanks

Mahesh.c

[former_member229109]

Hello mahesh chedde ,

-> do you mean ?

• Run on the node pisdbprd to reset DBM Server Shared Memory by executing the following command::

/sapdb/PIS/pgm/dbmshm DELETE /sapdb/data PIS

• After that you was able to connect to the database & restart it online:

dbmcli -d PIS -n pisdbprd -u control, <control>

< enter>

< you didn't get the error any more >

dbmcli on PIS> …

This, however, does not fix the cause of the situation.

Please don't recommend it as solution.

It's workaround to fix the problem with corrupted old Shared Memory files.

-> The question - 'Why you got the corrupted DBM Server Shared Memory files?' - still left open. The old Shared Memory files & more details needed for the further analysis of the problem.

You need to stop all DBM servers of an affected DB then remove its shi and shm files. Then the problem should be "solved".

Please inform us each time you have to remove the DBM Server Shared Memory, to resolve a problem. If the case unknown, it has to be checked by database developers.

It's good to find out how often & on which platforms this error occurs. Please provide platform information, the files <indep_data>/wrk/<dbname>.dbm.shi, <indep_data>/wrk/<dbname>.dbm.shm, <rundirectory>/dbm.prt and mount options type of file system for the <indep_data>/wrk directory.

Thank you and best regards, Natalia Khlopina