5 Replies Latest reply: Jun 5, 2014 3:06 AM by Hemant K Chitale RSS

    Standby database with Slow Internet connection

    2646875

      Please Advice i have oracle 11g r2 database with two physical standby databases one of them using LAN fiber optic connection with primary database  and another one using

      WAN with 2 m speed connection so when occur disconnect in WAN archiveslogs not ship

      to standby database and archives log have accumulate until WAN connection back after that the primary database have loaded until all arcgives log applied in standby database .

      so please find in below Dataguard setting

      the primary database run with good Memory and CPU under Linux

      what is the best practice for dataguard with WAN connection .

       

      log_archive_dest_1='location=/backup/archives mandatory reopen=300'

      *.log_archive_dest_2='service=ODFDB async noaffirm' LAN

      *.log_archive_dest_3='service=TTDB async noaffirm' WAN

      *.memory_target=8259633152

      and internet connection speed 2 M

      database 11.2.0.1

      operating system RedHat Linux 5.3 64bit

        • 1. Re: Standby database with Slow Internet connection
          Top.Gun

          If your network speed is not fast enough for dataguard then you need to get a faster network......

          • 2. Re: Standby database with Slow Internet connection
            2646875

            Hi top gun

            thank you for your  answer

            yes my network is slow but dataguard is working with delay so  i have bottleneck  performance in  primary database .

            how i can reduce archiving process performance  to met network speed

            • 3. Re: Standby database with Slow Internet connection
              Top.Gun

              "bottleneck  performance in  primary database"

              Are you saying the network speed is impacting the primary?

              Are you using maximum performance mode?

               

              "reduce archiving process performance  to met network speed"

              I don't think you want to reduce anything except the quantity of redo you need to send over the network.

              The only thing you can do is ensure you don't have too many indexes. Drop the ones you don't need, and this will reduce the redo that needs to be applied.


              • 4. Re: Standby database with Slow Internet connection
                2646875

                when archives have accumulate the primary database have loaded  while trying to ship archives log to standby database

                dataguard mode is maximum performance

                so find in below  alert log for primary database


                WARN: ARC5: Terminating ARCH (pid 31613) hung on a network operation

                ARC5: Error 16198 due to hung ARCH operation to 'TTDB'

                ARC5: Detected ARCH process failure

                ARC5: STARTING ARCH PROCESSES

                Thu Jun 05 08:47:24 2014

                ARC4 started with pid=66, OS id=31987

                ARC4: Archival started

                ARC5: STARTING ARCH PROCESSES COMPLETE

                Reclaiming FAL entry from dead process [pid 31613]

                Error 1031 received logging on to the standby

                Errors in file /database/app/oracle/diag/rdbms/ttdb/ttdb/trace/ttdb_arc9_31924.trc:

                ORA-01031: insufficient privileges

                FAL[server, ARC9]: Error 1031 creating remote archivelog file 'ttdb'

                FAL[server, ARC9]: FAL archive failed, see trace file.

                Errors in file /database/app/oracle/diag/rdbms/ttdb/ttdb/trace/ttdb_arc9_31924.trc:

                ORA-16055: FAL request rejected

                ARCH: FAL archive failed. Archiver continuing

                ORACLE Instance ttdb - Archival Error. Archiver continuing.

                Thu Jun 05 08:47:49 2014

                ARC4: Standby redo logfile selected for thread 1 sequence 19358 for destination LOG_ARCHIVE_DEST_3

                • 5. Re: Standby database with Slow Internet connection
                  Hemant K Chitale

                  >when archives have accumulate the primary database have loaded  while trying to ship archives log to standby database

                  No, the "load" on the primary doesn't increase much but it does keep retrying the send of the archivelogs.

                   

                  You have two different errors :

                  >WARN: ARC5: Terminating ARCH (pid 31613) hung on a network operation

                  and

                  >Error 1031 received logging on to the standby

                  Has there been a change at the primary or standby ?  The sys password changed and the password file not updated at the standby ?

                   

                  Hemant K Chitale