Oracle 18c changed the checkpointing in a physical standby database (with the side effect of a bug).
From: Martin Klier - Performing Databases GmbH <martin.klier_at_performing-db.com>
Date: Tue, 13 Aug 2019 15:25:12 +0200 (CEST)
Message-ID: <1403351687.4152.1565702712301.JavaMail.zimbra_at_performing-db.com>
Hi listers,
Date: Tue, 13 Aug 2019 15:25:12 +0200 (CEST)
Message-ID: <1403351687.4152.1565702712301.JavaMail.zimbra_at_performing-db.com>
Hi listers,
just in case you have a similar situation as I had:
- Oracle 18c
- Dataguard Physical Stby
- Log Sequences in question markled as "applied = yes" in v$archived_log
- but not applied on standby in reality
- Maybe Archive Gap that could not be resolved by FAL because FRA policies deleted the archives due to "applied = yes"
=> They changed the way how checkpoints work for MRP (now different to primary)
[ https://www.usn-it.de/2019/08/13/oracle-database-18c-changed-mrp-checkpointing-from-log-boundary-to-time-based/ | https://www.usn-it.de/2019/08/13/oracle-database-18c-changed-mrp-checkpointing-from-log-boundary-to-time-based/ ]
tl/dr: Oracle Dev said: " In 18.1 onward, time-based media recovery checkpoint is the default style of checkpoint (it's enabled via the hidden parameter _time_based_rcv_ckpt_target=180) meaning that by default, media recovery will perform a checkpoint every 180 seconds, instead of at every log boundary. "
Stay consistent!
Martin
-- Martin Klier // Performing Databases GmbH Managing Partner // Senior DB Consultant Oracle ACE Director martin.klier_at_performing-db.com // https://www.performing-databases.com -- http://www.freelists.org/webpage/oracle-lReceived on Tue Aug 13 2019 - 15:25:12 CEST