Oracle process dumping during RMAN backup, but only when Veritas ODM in use

book

Article ID: 100048791

calendar_today

Updated On:

Description

Error Message

Oracle alert log reported:

2020-07-02T12:56:40.923582+02:00
Process termination requested for pid 62883 [source = rdbms], [info = 2] [request issued by pid: 22305, uid: 778]
2020-07-02T13:23:03.327088+02:00
Running KSFV I/O slave I601 os pid=20694
2020-07-02T13:23:11.770442+02:00
ALTER SYSTEM ARCHIVE LOG
2020-07-02T13:23:11.832116+02:00
Thread 1 advanced to log sequence 2279 (LGWR switch)
Current log# 4 seq# 2279 mem# 0: /srv/oracle/gabome/data/GABOME/onlinelog/o1_m
f_4_hcob5w0r_.log
Current log# 4 seq# 2279 mem# 1: /srv/oracle/gabome/flash/GABOME/onlinelog/o1_
mf_4_hcob5z38_.log
2020-07-02T13:23:11.951602+02:00
NET (PID:19028): Archived Log entry 2276 added for T-1.S-2278 ID 0x195c6189 LAD:1
2020-07-02T13:23:25.579501+02:00
Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x0] [PC:0x0,

.

Errors in file /srv/oracle/gabome/admin/diag/rdbms/gabome/gabome/trace/gabome_or
a_20692.trc (incident=97608):
ORA-07445: exception encountered: core dump [PC:0x0] [SIGSEGV] [ADDR:0x0] [PC:0x
0] [Address not mapped to object] []

The stack trace from the core dump was received and the stack trace assessed:

file core.87227
core.87227: ELF 64-bit LSB core file x86-64, version 1 (SYSV), SVR4-style, from 'oraclegabome (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))', real uid: 778, effective uid: 778, real gid: 19037, effective gid: 19037, execfn: '/opt/oracle/product/19/bin/oracle', platform: 'x86_64'

gdb ./oracle ./core.87227

.
Core was generated by `oraclegabome (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))'.
Program terminated with signal 6, Aborted.
#0 0x00007ff58306a377 in raise () from /lib64/libc.so.6
Missing separate debuginfos, use: debuginfo-install glibc-2.17-292.el7.x86_64 libaio-0.3.109-13.el7.x86_64 libgcc-4.8.5-39.el7.x86_64

(gdb) bt
#0 0x00007ff58306a377 in raise () from /lib64/libc.so.6
#1 0x00007ff58306ba68 in abort () from /lib64/libc.so.6
#2 0x0000000010d7f613 in skgdbgcra ()
#3 0x000000000b6648e5 in sksdbgcra ()
#4 0x00000000072c8495 in ksdbgcra ()
#5 0x00000000126b8f5e in ssexhd ()
#6 0x0000000004239bb1 in sslssSynchHdlr ()
#7 0x00000000042398fe in sslsshandler ()
#8
#9 0x0000000000000000 in ?? ()
#10 0x00000000014f1656 in ksfdodm_remove_dir ()
#11 0x0000000000e3c584 in ksfdddcb ()
#12 0x0000000010d34bda in skgfrsftwn ()
#13 0x0000000010d34eb5 in skgfrsftwn ()
#14 0x0000000010d34eb5 in skgfrsftwn ()
#15 0x0000000010d33263 in skgfrsfen ()
#16 0x0000000000e3c18e in ksfddeomf ()
#17 0x000000000e96c64c in krbRmEmptyOMF ()
#18 0x0000000003c4d6d6 in krbbRmAllEmptyOMF ()
#19 0x0000000003c4bbbb in krbbcc ()
#20 0x0000000003c2d3d6 in krbbpcint ()
#21 0x0000000003c2b54d in krbbpcCbkArgs ()
#22 0x0000000003c2b2b5 in krbbpcCbk ()
---Type to continue, or q


 

Cause

Since no Veritas functions were seen in the core's stack trace, the issue was referred to Oracle.

Resolution

Oracle referred customer to Oracle patch 31753589 and core dumps stopped taking place.

Issue/Introduction

Oracle process dumping during RMAN backup, but only when Veritas ODM in use. In the setup, VRTSodm 7.4.1.1900 and Oracle Database 19c were used