Hello,
I found strange things inside zabbix_agentd 1.1 on openbsd:
zabbix_agentd -p
...stripped...
vfs.dev.write[hda,operations] [m|ZBX_NOTSUPPORTED]
vm.memory.size[free] [u|760807424]
proc.num[inetd,,] [u|1]
Segmentation fault (core dumped)
elvira# zabbix_agentd -t 'system.cpu.util[all,system]'
system.cpu.util[all,system] [d|43.299858]
(this number is still same, no respect of cpu state)
elvira# echo 'system.cpu.util[all,system]' | nc elvira 10050
0.000000
(this number is still same..)
So there is bug somewhere in meminfo which kills zabbix_agentd and second, cpu.util returns another value from -t and another from runing process.. First problem was already discused here. It is in openbsd implementation of proc.mem.
Second I do not understand. Probably there is bug inside openbsd/cpu.c and return value is sometimes not set.
Any suggestions?
I found strange things inside zabbix_agentd 1.1 on openbsd:
zabbix_agentd -p
...stripped...
vfs.dev.write[hda,operations] [m|ZBX_NOTSUPPORTED]
vm.memory.size[free] [u|760807424]
proc.num[inetd,,] [u|1]
Segmentation fault (core dumped)
elvira# zabbix_agentd -t 'system.cpu.util[all,system]'
system.cpu.util[all,system] [d|43.299858]
(this number is still same, no respect of cpu state)
elvira# echo 'system.cpu.util[all,system]' | nc elvira 10050
0.000000
(this number is still same..)
So there is bug somewhere in meminfo which kills zabbix_agentd and second, cpu.util returns another value from -t and another from runing process.. First problem was already discused here. It is in openbsd implementation of proc.mem.
Second I do not understand. Probably there is bug inside openbsd/cpu.c and return value is sometimes not set.
Any suggestions?