Hello,
i have host (storage) with two controllers (because redundancy), in host settings i created two interfaces (two IP adresses) - i noticed that one interface should be identified as default, i reading data via snmpv3.
How it works in zabbix when default interface is DOWN ? if im correct it should switch to secondary interface and read data from it, actually when it happens instead of switch to secondary interface, host just saying that there is timeout while connecting to interface (it is the default one).
While this situation im not recieving data for graphs, is there way how configure switchover of interfaces ? f.e. if one interface is down it will swiftly begin use the secondary interface and read data ?
Together with this behaving and i thing that is caused by symptomps wrote above, sometimes the unreachable pollers and poller collectors suddenly raise their load to maximum and zabbix queue also suddenly raise - am i correct that reason is that coulnd be interface for snmp reading is down and because zabbix cant reach it ?
Second issue is that when switchover happen the instance (last octet) of OID in items change for example from 1.1.1.1.1 to 1.1.1.1.2, (OID 1.1.1.1.1 is for reading through controller 1 and OID 1.1.1.1.2 is for reading through controller 2)
Do you have idea how solve this issue ? my idea is reading simultaneously from both controllers, write simultaneously data to graphs so always at least one source (controller) will be reachable, but what about unreachable pollers - i mean that it could be unwanted load of unreachable pollers ...who knows.
thank you for any idea
i have host (storage) with two controllers (because redundancy), in host settings i created two interfaces (two IP adresses) - i noticed that one interface should be identified as default, i reading data via snmpv3.
How it works in zabbix when default interface is DOWN ? if im correct it should switch to secondary interface and read data from it, actually when it happens instead of switch to secondary interface, host just saying that there is timeout while connecting to interface (it is the default one).
While this situation im not recieving data for graphs, is there way how configure switchover of interfaces ? f.e. if one interface is down it will swiftly begin use the secondary interface and read data ?
Together with this behaving and i thing that is caused by symptomps wrote above, sometimes the unreachable pollers and poller collectors suddenly raise their load to maximum and zabbix queue also suddenly raise - am i correct that reason is that coulnd be interface for snmp reading is down and because zabbix cant reach it ?
Second issue is that when switchover happen the instance (last octet) of OID in items change for example from 1.1.1.1.1 to 1.1.1.1.2, (OID 1.1.1.1.1 is for reading through controller 1 and OID 1.1.1.1.2 is for reading through controller 2)
Do you have idea how solve this issue ? my idea is reading simultaneously from both controllers, write simultaneously data to graphs so always at least one source (controller) will be reachable, but what about unreachable pollers - i mean that it could be unwanted load of unreachable pollers ...who knows.
thank you for any idea
Attached Files