[Glass] GC
Dale Henrichs via Glass
glass at lists.gemtalksystems.com
Fri Jan 16 09:52:33 PST 2015
Hello Leo,
If I understand you are trying to reduce the amount of persistent
garbage that gets produced while running reports and right now, you are
looking at the possible dead as a measure and you are wondering if there
is a different/better way. Is that right?
If you take a look at the `SystemRepository fileSizeReport` before and
after a report is generated, you should be able to get an idea about how
much disk space is consumed when you run the report. Unfortunately this
is not an exact science since some of the persistent objects created
during report generation may not be flushed to disk yet ... Running
`System startCheckpointSync` and then `fileSizeReport` will probably
give you a better measure of the disk space consumed by the operation.
`fileSizeReport` produces a string, so if you want to get your hands on
the real numbers then take a look at the implementation of
`fileSizeReport` to see where the numbers come from...
Dale
On 01/16/2015 04:07 AM, Leonardo Andres De Marco via Glass wrote:
> Hi all!
> Happy new year!
>
> We have a GC script that run 2 times a day, but sometimes when we
> needed, we just run on demand (for example, when the system perform a
> lot of reports).
>
> We are working on make the reports better to no generate much garbage,
> but the thing right now is how can we know how much garbage exist;
> some instructions to know if we have to run the gc or not?
>
> When gc script finish we can read something like this:
> "233058 possible dead objects, occupying approximately 20.975.220
> bytes, may be reclaimed." and we know for the bytes amount if the
> garbage was the performance problem or not.... thats ok?
>
> thnks!
> Leo
>
>
> _______________________________________________
> Glass mailing list
> Glass at lists.gemtalksystems.com
> http://lists.gemtalksystems.com/mailman/listinfo/glass
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/private/glass/attachments/20150116/2be229e4/attachment.html>
More information about the Glass
mailing list