Hi,
I've found 2 possible bugs.
possible bug #1
I've migrated 2 zabbix systems to a zabbix proxy based system (based on the latest trunk on svn) where 1 system is the zabbix server and 2 proxy's are relaying data.
When adding a couple of new checks the snmp_oid and the snmp_community columns were empty, because of this zabbix added h.proxy_hostid=0 in the query when it queries the zabbix items table for the check which was requested, because i was using a proxy the query was returning 0 rows, so zabbix thinks the item doesn't exist.
possible bug #2 (not really a bug but more a design thingie)
When defining a proxy, the proxy name is used to identify the remote proxy, so Hostname parameter in zabbix_proxy.conf equals the name of the proxy.
This seems unlogical to me as a proxy is just pointing to a host, maybe it's better to remove the proxy name field from the web interface and just let someone point to a host as proxy, then automatically use the hostname as the name of the proxy....
I've found 2 possible bugs.
possible bug #1
I've migrated 2 zabbix systems to a zabbix proxy based system (based on the latest trunk on svn) where 1 system is the zabbix server and 2 proxy's are relaying data.
When adding a couple of new checks the snmp_oid and the snmp_community columns were empty, because of this zabbix added h.proxy_hostid=0 in the query when it queries the zabbix items table for the check which was requested, because i was using a proxy the query was returning 0 rows, so zabbix thinks the item doesn't exist.
possible bug #2 (not really a bug but more a design thingie)
When defining a proxy, the proxy name is used to identify the remote proxy, so Hostname parameter in zabbix_proxy.conf equals the name of the proxy.
This seems unlogical to me as a proxy is just pointing to a host, maybe it's better to remove the proxy name field from the web interface and just let someone point to a host as proxy, then automatically use the hostname as the name of the proxy....