Hi Matt, I'll put together a small document tonight that covers what I remember about ganglia in the windows world and cygwin limitations and "good config practice" when doing unicast.
I also have a simple patch to implement cygwin cpu count that I will forward tonight. I remember a thread about this a while ago and I'm surprised that cygwin/gmetric.c/cpu_num_func() looks like the original one. - Richard -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Matthew Chambers Sent: Saturday, 25 August 2007 1:36 AM To: 'Richard'; ganglia-developers@lists.sourceforge.net Subject: Re: [Ganglia-developers] [PROPOSAL] Building Gmond on WindowsDoc... Hi Richard, It seems to me then that these are some additional things to note other than how to compile for Cygwin: the current broken behavior of Cygwin with multicast, that unicast behaves better but still quirky, and a sample gmond.conf that also disables the metrics which the Cygwin build doesn't currently support (like loadavg). I also noticed that the number of processors metric is broken on my dual core machine (it reports a single processor), which seems like something that Cygwin should be able to support without a problem. -Matt > -----Original Message----- > From: Richard [mailto:[EMAIL PROTECTED] > Sent: Thursday, August 23, 2007 11:17 PM > To: Matthew Chambers; ganglia-developers@lists.sourceforge.net > Subject: RE: [Ganglia-developers] [PROPOSAL] Building Gmond on > WindowsDoc... > > Ah yes, cygwin. > > As will be obvious to people running the cygwin gmond, multicast is > not supported and does not work. Multicast? Just say no. > > And as you discovered, there are also a few little things to be > done for unicast (or maybe just cygwin unicast). I do this: > > udp_send_channel { > host = 10.200.28.50 # Don't use 127.0.0.1 > port = 8649 > } > > udp_recv_channel { > port = 8649 > } > > That is, the headnode explicitly sends its own data to itself or else > you won't see it. > > As this means a different gmond.conf for every windows > host, we did a hack where the udp_send_channel host (the headnode) is > overridden (changed) if a certain registry key is present, and our > installation > process sets the registry key to be the real headnode hostname or IP. > My hack is too shameful to publish. > > regards, > richard > > ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.5.484 / Virus Database: 269.12.4/969 - Release Date: 23/08/2007 4:04 PM ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers
