Re: Power HA and Journaling.

A bit more detailed...

write Myrecord;

With FRCRATIO(*NONE) and no journaling
- 1 soft write to the file

with FRCRATIO(*NONE) and journaling w/o commit
- 1 soft write to the file
- 1 hard write to the journal

with FRCRATIO(*NONE) and journaling with commit
- 1 soft write to the file
- 1 soft write to the journal

with FRCRATIO(1) and journaling with or w/o commit
- 1 hard write to the file
- 1 hard write to the journal

with FRCRATIO(*NONE) and journaling w/o commit and #42 HA Journal
Performance
- 1 soft write to the file
- 1 soft write to the journal





On Thu, Jan 29, 2015 at 4:42 PM, Buck Calabro <kc2hiz@xxxxxxxxx> wrote:

On 1/29/2015 2:52 PM, Horn, Jim wrote:

I understand the performance penalty, but
I guess somehow bundling journal writes (not making them when they

happen)

is different from not writing data changes when they happen?
Still baffles me. I guess IBM just keeps it straight. The more it's

like

magic........

If I may...

With FRCRATIO(1), a database write looks like this:
INSERT INTO...
call database I/O API
wait for disk to acknowledge the write
return to INSERT

With journaling, the same transaction looks like this:
INSERT INTO...
call database I/O API
don't wait
call journal API, which runs in different process
don't wait
return to INSERT

So, while journaling seems to do an extra thing, it does it in parallel
with the rest of the I/O. I think the key word here is asynchronous.

--
--buck

'I had nothing to offer anybody except my own confusion' - Jack Kerouac
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.