Update: After letting this sit for a while and run, it appears as though the server is indeed receiving data from the OS X client, but it's receiving 10 minutes worth of data in one giant bulk update. I have "Interval" set to 60 in the configuration and the data being received does appear to be honoring that.
I've also noticed the timestamp that collectd on OS X is reporting is in UTC while on CentOS it is not. I'm not sure if that has anything to do with it or not. On Thu, Oct 1, 2009 at 12:47 PM, Tim Henrich <[email protected]> wrote: > I've got collectd working correctly from one CentOS system to another > CentOS server using the network, load, and perl plugin (on the server). > > I set up collectd on OS X 10.5.8 with the same configuration, started > collectd with -f, and watch the logs grab the load, and see: > > [2009-10-01 12:14:24] plugin_read_thread: Handling `load'. > [2009-10-01 12:14:24] plugin_dispatch_values: time = 1254413664; interval = > 60; host = mbp; plugin = load; plugin_instance = ; type = load; > type_instance = ; > [2009-10-01 12:14:24] uc_update: mbp/load/load: ds[0] = 0.750977 > [2009-10-01 12:14:24] uc_update: mbp/load/load: ds[1] = 0.653320 > [2009-10-01 12:14:24] uc_update: mbp/load/load: ds[2] = 0.632812 > [2009-10-01 12:14:24] plugin: plugin_write: Writing values via network. > [2009-10-01 12:14:24] plugin_read_thread: Effective interval of the load > plugin is 60.000000000. > [2009-10-01 12:14:24] plugin_read_thread: Next read of the load plugin at > 1254413724.828096000. > > On the server, I see the logs flying by from the CentOS client sending > data, but never see anything from the OS X client. > > After doing Control + C: > > ^C[2009-10-01 12:14:44] return (0); > [2009-10-01 12:14:44] Exiting normally. > [2009-10-01 12:14:44] collectd: Stopping 5 read threads. > [2009-10-01 12:14:44] plugin: stop_read_threads: Signalling `read_cond' > [2009-10-01 12:14:44] network plugin: flush_buffer: send_buffer_fill = 128 > [2009-10-01 12:14:44] network plugin: network_send_buffer: buffer_len = 128 > [2009-10-01 12:14:44] network plugin: sockent_destroy (se = 0x0); > > At this point I see the server getting the values, but only after the > client has been terminated. > > Both client and server have been configured with --enable-debug, but this > is as much information as I get. > > The only other clues are in /var/log/system.log on the client which says: > > Oct 1 12:16:13 mbp collectd[4905]: Libgcrypt warning: missing > initialization - please fix the application > > However, I've removed the encryption parameters from the network plugin and > the outcome is still the same. > > Any thoughts? > > -Tim > > >
_______________________________________________ collectd mailing list [email protected] http://mailman.verplant.org/listinfo/collectd
