Re: Performance tuning

[Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx>]

There is a lot of discussion in the help text about how this sort of 
caching "may not be suitable for interactive jobs where single system 
recovery is the primary reason for using journaling".  But that by 
definition seems to mean that it's okay when you're in an HA environment.

On 12/3/2018 9:03 AM, Darren Strong wrote:
> This is also referred to as journal caching, if you're searching for
> information on it. Its activated with the JRNCACHE(*YES) function in
> CHGJRN, once you purchase the option. In summary, journal data is cached
> into memory before writing to disk in larger chunks, rather than making
> sure its written at every change. It makes a big difference on journaled
> file write/update performance.
>
>
>
>
>
> From:   "Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx>
> To:     midrange-l@xxxxxxxxxxxx
> Date:   12/02/2018 01:06 PM
> Subject:        Re: Performance tuning
> Sent by:        "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>
>
>
>
> Yes, we've already gone to SSD, Rob, so that particular ship has already
> sailed.  :)  However, I am unfamiliar with HA Journal Performance.  We
> use MIMIX; I wonder if this is something I should be looking into.
> Heck, for all I know we already have it, but it definitely sounds like
> something I should look into.
>
> On 11/30/2018 12:31 PM, Rob Berendt wrote:
> > Sometimes throwing money at hardware isn't always the answer.  Our
> > original foray into SSD's was a big disappointment.  Much less money for
> a
> > much larger return on investment would have been to buy the "HA Journal
> > Performance" option as part of SS1.  I daresay it will be a rare
> > consultant who can pinpoint WHY you might need that and predict how much
> > it will help.
> >
> >
> > Rob Berendt