[rrd-developers] rrdtool 1.3 cairo switch is done
Dylan Vanderhoof
DylanV at semaphore.com
Fri Jun 15 21:44:47 CEST 2007
Apparently mmap overrules the fadvise functionality.
When compiled on 2.4.32 w/ --disable-mmap, I get this:
denver ~ # time ./rrdtool fetch test.rrd AVERAGE
ERROR: setting POSIX_FADV_RANDOM on 'test.rrd': Success
real 0m0.016s
user 0m0.010s
sys 0m0.000s
Not surprising, since posix_fadvise() isn't supported on 2.4.X.
Not sure is that's considered a regression, or if --disable-mmap should
be considered unsupported on 2.4.X
-D
> -----Original Message-----
> From: Dylan Vanderhoof
> Sent: Friday, June 15, 2007 12:32 PM
> To: Bernhard Fischer
> Cc: rrd-developers at lists.oetiker.ch
> Subject: Re: [rrd-developers] rrdtool 1.3 cairo switch is done
>
>
> I figured that's why it wasn't the default, but removing it had no
> effect on rrdtool performance, I was just playing with options.
>
> I'm still seeing approximatly half the performance from trunk
> than I am
> 1.2.23 + fadvise. (For fetch. graph is about 1/3 the speed
> with trunk)
> I'll without mmap when I get a chance to see if that's part of the
> issue.
>
> (FWIW, this is on linux 2.4.32, I haven't had a chance to test it on a
> 2.6.X server yet)
>
> -D
>
> > -----Original Message-----
> > From: Bernhard Fischer [mailto:rep.dot.nop at gmail.com]
> > Sent: Friday, June 15, 2007 12:25 PM
> > To: Dylan Vanderhoof
> > Cc: rrd-developers at lists.oetiker.ch
> > Subject: Re: [rrd-developers] rrdtool 1.3 cairo switch is done
> >
> >
> > On Fri, Jun 15, 2007 at 10:50:24AM -0700, Dylan Vanderhoof wrote:
> > >Tried this on my test server. Generated 515 graphs of ~30
> days each.
> > >Trunk functions fine and the graphs look good, but its
> substantially
> > >slower than 1.2.23 w/ the fadvise patch. (trunk compiled
> > with mmap and
> > >directio)
> >
> > no. Don't use directio (unless all your rrd are on a RAM backed
> > filesystem)!
> >
>
> _______________________________________________
> rrd-developers mailing list
> rrd-developers at lists.oetiker.ch
> https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers
>
More information about the rrd-developers
mailing list