Hi everyone, I have two servers running icinga and up until now each ran their own ido2db and icinga-web installation. After some research I found out I only need the icinga part running on one system and the ido2db can run centralized. Since these were different servers in different locations, I created an SSH tunnel from the non-ido2db machine effictively forwarding port 5668 on localhost to my remote host.
When I test this it works like a charm: I can manually connect to the port on both systems and starting icinga on both hosts and ido2db on the central node yields a nice green overview of all services in Icinga Web. However, after sometime, one node always drops out (about 10-15 min I guess). The only way to reconnect it then is restarting all parts of the system. Sometimes it work, sometimes I have to try multiple times. Also the instance does not reconnect by itself. The port forward is still open and working but that's not even it: Sometimes the localhost fails as well. >From the documentation I gathered that what I was trying to achieve is actually possible (centralising my monitoring while still running icinga itself separately). Here is my idomod of the remote server (Ubuntu 12.04): instance_name=jupiter output_type=tcpsocket output=127.0.0.1 tcp_port=5668 use_ssl=0 output_buffer_items=5000 buffer_file=/var/lib/icinga/idomod.tmp file_rotation_interval=14400 file_rotation_timeout=60 reconnect_interval=15 reconnect_warning_interval=15 data_processing_options=67108669 config_output_options=2 dump_customvar_status=0 debug_level=-1 debug_verbosity=2 debug_file=/var/log/icinga/idomod.debug max_debug_file_size=100000000 And here is idomod of the server running ido2db (ArchLinux): instance_name=default output_type=tcpsocket output=127.0.0.1 tcp_port=5668 use_ssl=0 output_buffer_items=5000 buffer_file=/run/idomod.tmp file_rotation_interval=14400 file_rotation_timeout=60 reconnect_interval=15 reconnect_warning_interval=15 data_processing_options=67108669 config_output_options=2 debug_level=-1 debug_verbosity=2 debug_file=/var/log/icinga/idomod.debug max_debug_file_size=100000000 Note that I adjusted the instance_name to be different on both systems. This is my tunnel: ssh -q -N -L 5668:localhost:5668 tunnel@myserver And here is my ido2db configuration: lock_file=/run/ido2db.pid ido2db_user=icinga ido2db_group=icinga socket_type=tcp socket_name=/var/spool/icinga/ido2db.sock socket_perm=0755 tcp_port=5668 use_ssl=0 db_servertype=mysql db_host=localhost db_port=3306 db_name=icinga db_prefix=icinga_ db_user=***** db_pass=***** max_timedevents_age=60 max_systemcommands_age=1440 max_servicechecks_age=1440 max_hostchecks_age=1440 max_eventhandlers_age=10080 max_externalcommands_age=10080 max_logentries_age=44640 max_acknowledgements_age=44640 max_notifications_age=44640 max_contactnotifications_age=44640 max_contactnotificationmethods_age=44640 clean_realtime_tables_on_core_startup=1 clean_config_tables_on_core_startup=1 trim_db_interval=3600 housekeeping_thread_startup_delay=300 debug_level=-1 debug_verbosity=2 debug_file=/var/log/icinga/ido2db.debug max_debug_file_size=100000000 debug_readable_timestamp=0 oci_errors_to_syslog=1 oracle_trace_level=0 enable_sla=0 Has anyone experienced that? As a sidenote: On my localserver, idomod.debug gets spammed with the message "[1391644219.095614] [001.2] [pid=9460] idomod_broker_data() start" (multiple times per seconds). Could this be related? Or is this no problem? It seems wasteful and also seems to kind of increase CPU but I could be wrong. _______________________________________________ icinga-users mailing list [email protected] https://lists.icinga.org/mailman/listinfo/icinga-users
