[rrd-users] Out of memory restoring a big dump file

[Tobias Oetiker]

Today Karl Fischer wrote:

> Am 22.03.2011 14:01, schrieb Tobias Oetiker:
> > Hi Fabio,
> >
> > note that the restore code has been completely rewritten in 1.4 and
> > is much more memory efficient now ... use 1.4
> >
> > cheers
> > tobi
>
> uhh - didn't know that - interesting news.
> The 220 MB memory usage I mentioned was using 1.4.5
> I just tried it with 1.3.1 and it went right through the roof
> (I killed the process after it took 7 GB of ram ...)

the 1.3 version used a classic xml parser ... while 1.4 is now
using a pull parser and thus uses more somethig in the region of
the amount of memory occupied by the original rrd file.

if we did two passes of the input file, this could even be reduced
further

cheers
tobi

>
> - Karl
>
>

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland
http://it.oetiker.ch tobi at oetiker.ch ++41 62 775 9902 / sb: -9900