Open julianbrost opened 10 years ago
On 2014-08-04 10:48, Julian wrote:
Naemon takes multiple minutes to start with about 100 hosts and 2000 services with
check_period nopowersave
as defined in this config section:
Yes. This is a known issue. It basically boils down to the fact that get_next_valid_time() and get_next_invalid_time() iterate 300 times in order to find a suitable timerange (ie, a valid day where there actually is one). It's more or less insane, but that's the reason for it anyways.
I'm working on changing that, but coming up with a correct and clever algorithm is proving difficult, since we really don't want the daylight savings time bug to come back again.
/Andreas
Naemon takes multiple minutes to start with about 100 hosts and 2000 services with
check_period nopowersave
as defined in this config section:The debug log mostly contains the message
_get_matching_timerange()
, here's a short snippet of it, servicefoo
and hostbar
havecheck_period nopowersave
configured.Stracing the process gives loads of stats on
/etc/localtime
:GDB backtrace while startup:
Naemon version is 0.8.0 with the patch from commit ad800d8a0e14901bf4adc31231693ca0fd9fa9f2 applied, running on Debian wheezy.