Hello,
We are integrator of Huawei products in czech republic and trying to implement zabbix in our projects.
Im newbie in Zabbix at all, but i already learn some simple skils like - how add host, how walk OID, how make items, triggers and how make working alarms, thats clear.
I have made all of above for 215 servers and everything work fine, servers have "static" OIDs, but now i should make it for storage appliences (SNMP V3) and i have problem with items/OIDs.
The problem is that OIDs for storage appliace (Huawei Ocean Stor 5500 V5) can change itself by time goes by, i can interpret is on this OID:
It has 3 parts:
1.) first part
2.) second part
6 = 6 following numbers (48.46.48.66.46.49) this one number (6) is also static part, it cant change itself by time goes by
3.) last six numbers
48.46.48.66.46.49 -> this is "instance part" 4rd number can change itself randomly between 65/66 but whole OID is still belongs to same HW component.
How can i "learn" Zabbix to count with this changes ?
Additional info: Why the OID change itself randomly ? it is because monitored device is storage appliance which has two controllers, both are connected to same network (each has own IP) always only one controller is active (all snmp reads flow through active controller) and active/inactive status rotates between this two controllers.
When i created Host i added both IP of controller A & B into one host.
Example of changing OID´s:
OID: 1.3.6.1.4.1.34774.4.1.23.5.5.1.3.6.48.46.48.65.46.49 belongs to controller A
OID: 1.3.6.1.4.1.34774.4.1.23.5.5.1.3.6.48.46.48.66.46.49 belongs to controller B
Ufortunately we cant read from both controller at one time, there is always only one "Master controller" from which we read values (and OID), unfortunately by time goes by, "Master controller" can be changed automatically from controller A to controller B and vice versa. I have tried to solve it by creating doubled items, for example item for CPU1 temperature was doubled to (CPU1 temperature A & CPU1 temperature B - both items read value from CPU1 but in item A is OID with 65=controller A and in second item B is OID with 66= controller B) which seems like good idea because when switch of Master controller happen we still have one item which read values correctly, unfortunately the item which is "inactive" in that moment (item wich belongs to non-master controller at the moment) reporting error "NOT SUPPORTED" because this OID dont respond at that time, if i have only 2 items reporting this error it is ok but when more items reporting this error it shutdown whole Zabbix, but i dont know why ? standard period of reading is 300seconds but when appliance dont respond, how item will behave ? it is period will be shortened it may overload zabbix server and of course it can cause crash but it is only my deduction and im not sure if this solution is correct and effective so this is reason why i posted this question.
If anyone will have any question please ask me i will try hard to give you all needed informations.
Thank you very much.
Filip
We are integrator of Huawei products in czech republic and trying to implement zabbix in our projects.
Im newbie in Zabbix at all, but i already learn some simple skils like - how add host, how walk OID, how make items, triggers and how make working alarms, thats clear.
I have made all of above for 215 servers and everything work fine, servers have "static" OIDs, but now i should make it for storage appliences (SNMP V3) and i have problem with items/OIDs.
The problem is that OIDs for storage appliace (Huawei Ocean Stor 5500 V5) can change itself by time goes by, i can interpret is on this OID:
| 1.3.6.1.4.1.34774.4.1.23.5.5.1.3.6.48.46.48.66.46.49 |
1.) first part
| 1.3.6.1.4.1.34774.4.1.23.5.5.1.3. -> this is static part, it cant change itself by time goes by |
6 = 6 following numbers (48.46.48.66.46.49) this one number (6) is also static part, it cant change itself by time goes by
3.) last six numbers
48.46.48.66.46.49 -> this is "instance part" 4rd number can change itself randomly between 65/66 but whole OID is still belongs to same HW component.
How can i "learn" Zabbix to count with this changes ?
Additional info: Why the OID change itself randomly ? it is because monitored device is storage appliance which has two controllers, both are connected to same network (each has own IP) always only one controller is active (all snmp reads flow through active controller) and active/inactive status rotates between this two controllers.
When i created Host i added both IP of controller A & B into one host.
Example of changing OID´s:
OID: 1.3.6.1.4.1.34774.4.1.23.5.5.1.3.6.48.46.48.65.46.49 belongs to controller A
OID: 1.3.6.1.4.1.34774.4.1.23.5.5.1.3.6.48.46.48.66.46.49 belongs to controller B
Ufortunately we cant read from both controller at one time, there is always only one "Master controller" from which we read values (and OID), unfortunately by time goes by, "Master controller" can be changed automatically from controller A to controller B and vice versa. I have tried to solve it by creating doubled items, for example item for CPU1 temperature was doubled to (CPU1 temperature A & CPU1 temperature B - both items read value from CPU1 but in item A is OID with 65=controller A and in second item B is OID with 66= controller B) which seems like good idea because when switch of Master controller happen we still have one item which read values correctly, unfortunately the item which is "inactive" in that moment (item wich belongs to non-master controller at the moment) reporting error "NOT SUPPORTED" because this OID dont respond at that time, if i have only 2 items reporting this error it is ok but when more items reporting this error it shutdown whole Zabbix, but i dont know why ? standard period of reading is 300seconds but when appliance dont respond, how item will behave ? it is period will be shortened it may overload zabbix server and of course it can cause crash but it is only my deduction and im not sure if this solution is correct and effective so this is reason why i posted this question.
If anyone will have any question please ask me i will try hard to give you all needed informations.
Thank you very much.
Filip