Opened 16 years ago
Last modified 16 years ago
#2 accepted task
check Host alive genesis
Reported by: | maarten | Owned by: | maarten |
---|---|---|---|
Priority: | major | Milestone: | Service checks collection no.1 |
Component: | Check plugins | Version: | pre-exodus |
Keywords: | check host alive icmp snmp | Cc: |
Description
Proper host check for wleiden configuration based on genesis.
specific points:
- snmp not running on all hosts
- all ip's are bind to an interface
Check ICMP to all host ip's with result of 1 up is host up.
Note by DWG:
SNMP uptime check is betrouwbaarder - want dan weet je dat niet alleen de kernel up is - maar ook userland
Note:
See TracTickets
for help on using tickets.
the check_icmp on all interfaces workes nice...
when exodus configuration is roled out there is a local loopback ip that is always available. checking only this ip will speed up the host_alive check!
One problem is that the host down check is faster then the lvrouted deamon to reroute when a host is down.
If this happens you see a bunch of nodes go down and then recover.
This should be finetuned so nagios waits a bit longer before stressing out...
also the parents directive of the hosts could assist in some situations.