The bug I noticed is only in the web-frontend meta view. If the scalable is set to off then the grid summary table and plots do not show up on the top of the page anymore. It just shows up in the middle of the list of clusters with a blank name. On the top, and next to the cluster selection box, it says "XXX Grid", using the name of one of our clusters for XXX.
My guess is that this is happening because there is no grid name in the xml data anymore and the web-frontend assumes it will get a grid name. This is why I suggested that gmetad should still write the grid tag out, but ignore it on input, so the grid name is always present, for the web-frontend and other gmetads to use if they like. It is up to the higher gmetad to decide if it will be scalable or not and ignore the grid tag and its authority location. The source gmetad can offer the scalable option but the higher gmetad can choose to ignore it and reproduce all rrds locally. Of course, it would also be nice if the source gmetad could report its grid name for a nicer tree layout, but somehow let the higher level gmetad know that it is not available for authority referrals, maybe because of firewall or other security issues. ~Jason On Mon, 2003-09-01 at 00:11, Federico Sacerdoti wrote: > On Friday, August 29, 2003, at 12:14 PM, Jason A. Smith wrote: > > > There is an inconsistency for scalability mode in the example > > gmetad.conf config file and the gmetad/conf.c source file which reads > > that config file. In the example config file, the option is called > > "scalability" but the source is looking for an option named "scalable". > > > Yes, this is changed in CVS. The sample config file should say > "scalable". > > > Also, if the scalable option is turned off then the meta view on the > > webfrontend gets a little screwed up. Maybe gmetad should always write > > the GRID tag, but scalable mode just causes it to ignore it if it reads > > it from a data source. > > > I've tested a very large cluster with the scalable option off and it > worked fine. Perhaps I did not stress the meta view enough to see the > bug you speak of. > > Gmetad is designed to ignore the grid tags with the "scalable=off" > mode. All clusters will seem "local" and RRDs will be kept for > everyone. What is the bug you saw? > > fds > > ~Jason > > > > > > -- > > /------------------------------------------------------------------\ > > | Jason A. Smith Email: [EMAIL PROTECTED] | > > | Atlas Computing Facility, Bldg. 510M Phone: (631)344-4226 | > > | Brookhaven National Lab, P.O. Box 5000 Fax: (631)344-7616 | > > | Upton, NY 11973-5000 | > > \------------------------------------------------------------------/ > > > > > > > > ------------------------------------------------------- > > This sf.net email is sponsored by:ThinkGeek > > Welcome to geek heaven. > > http://thinkgeek.com/sf > > _______________________________________________ > > Ganglia-developers mailing list > > Ganglia-developers@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/ganglia-developers > > > Federico > > Rocks Cluster Group, San Diego Supercomputing Center, CA -- /------------------------------------------------------------------\ | Jason A. Smith Email: [EMAIL PROTECTED] | | Atlas Computing Facility, Bldg. 510M Phone: (631)344-4226 | | Brookhaven National Lab, P.O. Box 5000 Fax: (631)344-7616 | | Upton, NY 11973-5000 | \------------------------------------------------------------------/
