3 Replies Latest reply on Feb 18, 2016 5:58 AM by CloudDB

    facing segfault issue on linux

    SVM 1225

      Dear All,

       

      We are having database 11gR2 deployed on redhat linux.

       

      OS Version: Red Hat Enterprise Linux Server release 5.8 (Tikanga)

      Database Version : Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production

       

      We are getting below error messages in /var/log/messages.

       

      Feb 7 01:05:18 xyz kernel: oracle[15402]: segfault at 0000003463400880 rip 0000003463400880 rsp 00007fff1b880c38 error 14

      Feb 7 01:05:18 xyz kernel: oracle[15479]: segfault at 0000003463400880 rip 0000003463400880 rsp 00007fff94c94358 error 14

      Feb 7 01:05:18 xyz kernel: oracle[15446]: segfault at 0000003463400880 rip 0000003463400880 rsp 00007fffc3c2c3c8 error  14

       

      The performance of the database is getting slow.


      Can anyone share the possible solutions or suggestions to resolve the issue?


      Regards,

      SVM1225

        • 1. Re: facing segfault issue on linux
          oralicious

          you have 2 issues, you have segmentation faults you need to get the answer to and you have a slow database.   Investigate the 2 separately, are they related? maybe, maybe not. 

           

          Id start with tracing whats generating the seg faults.  Do that first.  Many possible reasons with many notes out on how to do same out there.   If oracle is generating them then youll get a dump which you should upload to oracle support as theyre not meant for simple folk like you or I.

           

          check your cdump trace folder for same.  if theres a dump file that matches the same time as the segfault messages then you're on the right road.   may or may not be related to your slow database. 

          1 person found this helpful
          • 2. Re: facing segfault issue on linux
            Vsevolod Afanassiev

            In this message "Feb 07 01:05:18  xyz kernel: oracle[15402]  segfault at"   word "oracle" means the process owner. So it doesn't have to be the database, it could be a shell script or something else running as UNIX user "oracle". Check alert.log - if it shows ORA-00600, ORA-07445, or similar errors at the same time as /var/log/messages then it is database issue.

            1 person found this helpful