archiver

[jok llamera]

hi fellas,

Im sure one of you here is an Oracle DBA, Im using HPUX 11.0 and I have a question when writing a data from redolog(RAW) to the archivelog files(Filesystem). Does it bypasses the oracle buffer cache or not?

Thanks,
joks

[Carlos Fernandez Riera]

And the answer is NO.


You can check it with tusc:

tusc -f -o /tmp/tusc.log -a -s open,read,write -p PID

PID is PID of ora_arch_$ORACLE_SID

It seem to open redolog file for read only, open a new redolog_on_disk file and write while read.

Regards.

[Andreas D. Skjervold]

Hi

The archiver will use the Redo Log Buffer.

The archiver will:
1) read the control file to find any unarchived logs,
2) open online redo log members to be read,
3) allocate redo log buffers,
4) read the online redo log,
5) fill redo log buffers,
6) write to the archive files,
7) update control file with new information,
8) starts the loop again.

Andreas

[Ryan Kogelheide]

Andreas,

That's interesting. I've obviously misunderstood something about how the archiver works.

What I always imagined is that it would read and edit the control file while doing a simple OS copy.

Why does it need to work with redo buffers?

Ryan