Re: Archive Log Size

The block size of redo logs is the sector size of discs, which is 512 bytes (historically) and the archived redo log is an image copy of the redo log up to the highwater mark (unless you've enabled compression on archiving). New very large discs may have a sector size of 4K so Oracle introduced the option for using 4K redo blocks (which could mean a much larger redo wastage as Oracle doesn't backfill partially used redo blocks after writing them) or for emulating 512 byte blocks (which means potentially re-reading a 4KB block to fill the next 512 bytes on a write).

For a very long time archived redo logs have tended to be smaller than online redo logs, but initially that was because of things like "archive_lag_target" being set to force log file switches by time rather than waiting for filled logs (and only the used portion would be archived); but when Oracle introduced the whole multiple log buffer thing it became common for archived redo logs to be smaller (sometimes much smaller) than the online logs even when there was nothing forcing log file switches, See https://www.red-gate.com/simple-talk/sql/oracle/are-your-oracle-archived-log-files-much-too-small/

I think arcive_lag_target = 0 might be one of those funny zeroes that means use the default value (viz: 900).

Regards
Jonathan Lewis

On Tue, 9 Mar 2021 at 20:14, Chris Taylor <christopherdtaylor1994_at_gmail.com> wrote:

> I thought it was normal for archivelogs to be smaller than redo logs. The
> block size on archivelogs is always 512 whereas the block size for redo
> logs is 8k (by default) I think.
>
> I didn't think archivelogs would ever be the same size as redo logs. (I'm
> pretty sure I've never seen archivelogs be the same size as redologs)
>
> Our redologs are 20G and we have archivelogs ranging from 89MB to 18G
> (upper limit) with archive_lag_target target set on the primary to 0
> (though I thought it was 900 on the primary).
>
> Chris
>
> Chris
>
>
>

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