After Switch /Failover Primary Instance Open hangs Because Of SRL Cleanup (Doc ID 2710349.1)

To Bottom

In this Document

APPLIES TO:

Oracle Database - Enterprise Edition - Version 12.2.0.1 and later
Information in this document applies to any platform.

SYMPTOMS

After Switchover or Failover the Primary database open Hangs Clearing SRL

2020-08-22T16:22:20.172813+02:00
TT05 (PID:339034): SRL selected for T-2.S-7297 for LAD:2
2020-08-22T16:22:21.167782+02:00
TT04 (PID:339013): Sleep 10 seconds and then try to clear SRLs in 3 time(s)
2020-08-22T16:22:31.171341+02:00
TT04 (PID:339013): Sleep 20 seconds and then try to clear SRLs in 4 time(s)
2020-08-22T16:22:51.174015+02:00
TT04 (PID:339013): Sleep 40 seconds and then try to clear SRLs in 5 time(s)
2020-08-22T16:23:31.177369+02:00
TT04 (PID:339013): Sleep 80 seconds and then try to clear SRLs in 6 time(s)
2020-08-22T16:24:48.513050+02:00

Issue is also reported when converting to Snapshot database where the database open would hang and the hang analyze would show

Chain 1

is waiting for 'row cache lock' with wait info:

current sql: update /*+ rule */ undo$ set name=:2,file#=:3,block#=:4,status$=:5,user#=:6,undosqn=:7,xactsqn=:8,scnbas=:9,scnwrp=:10,inst#=:11,ts#=:12,spare1=:13 where us#=:1

short stack: ksedsts()+426<-ksdxfstk()+58<-ksdxcb()+872<-sspuser()+200<-__sighandler()<-epoll_wait()+19<-ipcor_wseti_wait()+94<-ipclw_wait()+520<-ksxpwait_ipclw()+920<-ksxpwait_int()+155<-ksxpwait()+319<-ksliwat()+2321<-kslwaitctx()+188<-kjusuc()+5580<-ksipgetctxia()+1968<-ksixpgetia()+124<-kqrInstanceLock()+667<-kqrGetClusterLock()+2547<-kqrLockPo()+407<-kqrpre2()+1106<-kqrpre1()+40<-ktuGetRowCache1()+82<-ktuGetUsegDba()+589<-ktuGetCommitTimesMain_Int()+626<-ktuGetCommitTimesMain()+542<-ktuGetCommitTimes()+149<-ktbgcl1()+2614

CAUSE

The issue has been repored in Unpublished bug BUG 31747989

Unpublished Bug 30159581 ALTER NEW PRIMARY OPEN HANG AFTER SWITCHOVER which is superseded by Unpublished Bug fix BUG 31747989 TRACKING BUG FOR DEPENDENT FIXES OF FIRST-PASS TRANSACTION RECOVERY

Unpublished BUG 30941790 DATA GUARD FAILOVER HANGS ON DATABASE OPEN this also has been closed as Duplicate of same Unpublished BUG 31747989 - TRACKING BUG FOR DEPENDENT FIXES OF FIRST-PASS TRANSACTION RECOVERY

SOLUTION

Workaround is to use :-

_min_undosegs_for_parallel_fptr=0

alter system set "_min_undosegs_for_parallel_fptr"=0 scope=both sid='*' ;

Restart the db

Or

Request an one off Patch for Bug 31747989 or Check for availability of one off Patch Patch 31747989