> gmetad needs to have info on the data sources from the > subclusters. Theoretically we could have two paths: > 1) start gmetad in post_server_install for the case of no > other data sources, > 2) start gmetad in post_install if there are other data > sources (per image > gmond config). > > Because there is no need for gmetad until post_install, I > chose to have only > one path: 2). It seems to be simpler.
I would prefer to go with option 1) if possible. The ganglia page will be displaying wrong information between post_server_install and post_install. I think it is safe to assume that after you pressed "Install OSCAR Server Packages" that the packages you have selected to be correctly installed (at least) on the server. > You can select in the central ganglia configuration whether > there should be a > per image config or not. Default is no per image config. In > that case the > central ganglia config is used for all images and the user is > not asked to > configure anything after building an image. > > Unfortunately there is no moment before building an image > when OSCAR _knows_ > how many images will really be there. This is why with the > per image config it > will ask for configuration right after you've built another image. I did not configure the Ganglia package in Step 2 (so should have no per-image configuration), yet after my only image was built, it prompted me to configure the package - is this a bug? Thanks, Bernard ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://ads.osdn.com/?ad_idv37&alloc_id865&op=click _______________________________________________ Oscar-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oscar-devel
