After a new installation of an Oracle 19 CDB+PDB database we always look at the alert.log.

And there were 2 anomalies:

1: ORA-00800:

ORA-00800: soft external error, arguments: [Set Priority Failed], [VKTM], [Check traces and OS configuration], [Check Oracle document and MOS notes], [] Error attempting to elevate VKTM’s priority: no further priority changes will be attempted for this process

ALERt.LOG

So the customer had to open a ticket with RHEL:

The latter already knew the problem:

  • Removing any software which enabled CPUQuota and/or CPUaccounting on systemd level (usually Insights), or remove the CPUQuota and accounting directives from systemd files.
    This solves ORA-00800 error without need of changing cpu.rt_runtime_us.
  • Why does Insights prevent Oracle grid services from starting
    https://access.redhat.com/solutions/6020561
  • Oracle Database Enterprise 19c fails to start with error: ORA-00800: soft external error, arguments: [Set Priority Failed], [VKTM]
    https://access.redhat.com/solutions/6991342



2: Prallel FPTR failed:

PDB(3):Undo initialization recovery: Parallel FPTR failed: start:28373 end:28380 diff:7 ms (0.0 seconds)

ALERt.LOG

support.oracle.com:

Workaround:

The default value for _min_undosegs_for_parallel_fptr is 100, which you can change to 0. This problem appears to be documented in the note below and fix included in oracle 20.0 version.

Bug 30159581 – A DB open hangs after switchover due to a detected deadlock ( Doc ID 30159581.8 )

Please implement work around solution as discussed in the ( Doc ID 30159581.8 )

Oracle DBA Life

yaitcon