Re: Redo for the Undo

Hello Lothar,
let's go one step back first.

How does the (ASH) wait chain look like when you experience this "log file switch (checkpoint incomplete)" issue? What's the redo/sec rate at this time? What's the DBWR doing/waiting on? Are the data files located in ASM or a filesystem (if yes, which one)?

There are several root causes why the DBWR can't keep up with the load ... so if the load is not very high (regarding the underlying infrastructure capabilities) ... you should dig into there.

Best Regards
Stefan Koehler

Independent Oracle performance consultant and researcher Website: www.soocs.de
Twitter: _at_OracleSK

> Lothar Flatz <l.flatz_at_bluewin.ch> hat am 18.12.2023 08:57 CET geschrieben:
>
> What we see are "log file switch (checkpoint incomplete) ". I first wanted to understand the basics before jumping in a conclusion.
> The "log file switch (checkpoint incomplete)" are always accompanied by big deletes (> 100 million rows).
> The deletes are a necessary but not sufficient criteria. Since we see big deletes without "log file switch (checkpoint incomplete)" waits.
> "log file switch (checkpoint incomplete) " means that DBWR can not keep up with LGWR. I wondered how that happens.

--
http://www.freelists.org/webpage/oracle-l