Hi Henrik, > from David's. One obvious difference between our systems is the memory > size; but we do share the fact that our set of RRD files does not fit in > the buffer cache.
one of the major points of this exercise is to make all 'hot' blocks of all rrds fit into memory. if they do not fit, then the performance gain will be much less ... the reason for this is, that for every time a few bytes are written, an entire block has to read back from disk, altered in memory and then written back to disk ... if the block is already in cache, it does not have to be brought back first, which makes things quite a bit faster ... by using fadvise in various ways we try to tell the system to only keep the blocks in memory that we need for updates ... cheers tobi -- Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten http://it.oetiker.ch [EMAIL PROTECTED] ++41 62 213 9902 _______________________________________________ rrd-developers mailing list rrd-developers@lists.oetiker.ch https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers