[collectd] perl plugin: counter values 2^31-1?

Hi all, I ran into this scenario with a perl write plugin: rhel5 i386, collectd 4.9.1 $ uname -a Linux myhost 2.6.18-92.el5PAE #1 SMP Tue Apr 29 13:31:02 EDT 2008 i686 i686 i386 GNU/Linux 1) A read plugin submits a counter value 2^31-1 2) When my perl write plugin sees value list, the value

Re: [collectd] [PATCH] Bugfix: correctly handle spurious wakeups in pthread_cond_timedwait

Hi Michael, On Wed, Feb 10, 2010 at 03:42:11PM +0100, Michael Stapelberg wrote: C99 also includes stdbool.h, which defines bool as an alias for _Bool. I know, but iIrc we've had problems with stdbool.h in the past. Some - arguably broken - libraries define true and false themselves. So I prefer

Re: [collectd] [PATCH] Initialize libsensors with NULL as parameter

Hi Michael, On Tue, Feb 09, 2010 at 05:30:16PM +0100, Michael Stapelberg wrote: The attached patch uses NULL for sensors_init. The downside of this is that the mechanism to reload the configuration file cannot be used any longer. I'm a bit concerned about backwards compatibility here:

Re: [collectd] write_http not working

Hi, On Fri, Feb 12, 2010 at 11:19:36AM -0700, Paul Sadauskas wrote: I pinged octo on IRC, hopefully this will be fixed soon. just wanted to let you know that I merged 4.8 into 4.9 and master, so the patch should be included in all maintained branches now. Thanks again :) Regards, -octo --

Re: [collectd] Collectd compatibility issue

Hi Ivaylo, On Fri, Feb 12, 2010 at 05:01:26PM -0800, Ivaylo Lenkov wrote: So my questions are: 1. is version 4.3.0 over-the-wire compatible with 4.6.3? Yes, all 4.* versions should be compatible to one another. In some occasions, the server (i.e. the machine *receiving*) must be *at least*

Re: [collectd] [PATCH] collectd: Use sensor labels (description) from lm-sensors

On Thu, 25 Feb 2010 23:44:11 +0100 Florian Forster o...@verplant.org wrote: I suggest to implement a new configuration option UseLabels. It defaults to false to now change the default behavior and when enabled uses the labels defined in the lm_sensors config. Does that sound reasonable?