Hi Ben: On Wed, Apr 23, 2008 at 4:01 PM, Ben Rockwood <[EMAIL PROTECTED]> wrote:
> The question I have for the list is... has anyone else run into this > bottleneck? If so, how did you solve the issue? I know several people > have multiple gmetad's out there, perhaps unnecessarily. CPU usage and > disk consumption are low, its just doing a lot of IO. My data/ dir is > only 27MB. I think with one gmetad, when you start monitoring hundreds to thousands of nodes, you'll run into the RRDtool I/O issue (of course depends on if you have extra "custom" metrics). It has been discussed numerous times (check archives for more details) and the most common solution is basically what you did -- put the rrd directory in tmpfs. I heard that the next version of RRDtool (1.3) is supposed to have some scalability and I/O improvements. It would be nice if someone who has tested the beta on a largish cluster could share their experiences. Cheers, Bernard ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
