snmpcollector , snmpget and snmp related probes

[MADANRAJ SIVAGNANAM]

We should use snmpcollector , snmpget and snmp related probes on robots in CA UIM  and so whether we should deploy these snmp related probes on robots in primary hub or secondary hub ?

what is the load on primary hub if confiured on primary hub ?


Regards
Amar

[Luc Christiaens]

snmpcollector can have a big impact on the robot/hub where it's installed.
It's strongly advised NOT to install it on the primary hub, better to install it on/under a secondary hub.

[MADANRAJ SIVAGNANAM]

Hi Luc ,

I need to add one more point here . In our environment , we are following like below mentioned robot.cfg file for configuring a robot . In this context , you can understand that we are running both the primary and secondary hubs simultaneously ,.  Now , you please let me know if I install snmpcollector on robot in secondary hub is it ok ??? 

<controller>
domain = NPCI_domain
robotip =
hubip = 10.1.112.134
first_probe_port = 48000
hub = CHECAUIMHUB01_Pri_hub
hubrobotname = CHECAUIMHUB01
hubport = 48002
access_0 = 0
access_1 = 1
access_2 = 2
access_3 = 3
access_4 = 4
ip_version = ipv4
local_ip_validation = no
strict_ip_binding = no
robotname = HYDBSATMSW02
secondary_domain = HYDCLDUIMPHUB01_Domain
secondary_hub = HYDCLDUIMPHUB01_hub
secondary_hubrobotname = HYDCLDUIMPHUB01
secondary_hubip = 192.168.63.239
secondary_hubport = 48002
robotip_alias = 10.219.132.16
secondary_robotip_alias = 10.219.132.16
loglevel = 0
logsize = 1024
autoremove = no
suspend_on_loopback_only = yes
config_locking = no
hub_update_interval = 900
temporary_hub_broadcast = yes
system_uptime_qos = no
set_qos_source = no
unmanaged_security = domain_locked
alarm_level_comfail_restart =
alarm_level_dispatch_error = major
alarm_level_max_restarts = major
alarm_level_start_error = major
alarm_level_timed_not_finished = warning
alarm_level_unregister = major
alarm_level_request_error = major
alarm_level_timed_error_return = warning
alarm_level_postinstall = no alarm
robot_mode = normal
</controller>

[Luc Christiaens]

Why did you use a secondary_domain that is not standard domain?
I believe that some queue/hub connections/commands are only possible in the same UIM domain (at least that was the rule some releases back) 
Do you have a visio or .jpg of your architecture?

[David Michel]

It's in the documentation
https://docops.ca.com/ca-unified-infrastructure-management-probes/ga/en/alphabetical-probe-articles/snmpcollector-snmp-data-monitoring/snmpcollector-snmp-data-monitoring-release-notes/install-snmpcollectorf

Installing the probe on the primary hub, or on a robot alone is not supported.

------------------------------
Support Engineer
Broadcom
------------------------------

[MADANRAJ SIVAGNANAM]

Hi Luc ,

Why did you use a secondary_domain that is not standard domain?
I believe that some queue/hub connections/commands are only possible in the same UIM domain (at least that was the rule some releases back)  

I did not understand the above lines given by you . we have 2 hubs that means what I am trying to convey is we are running both the hubs seperately and nimsoft robot watcher is running inside both the hubs simultaneosly . 


Please let me know what need to provided for you .
.jpg architecture ---> I did not get this . We are using CA UIM architecture as defined .


Regards
Amar

[Luc Christiaens]

From your settings I can see you have a different "domain" and "secondary_domain"
In 1 UIM environment you create normally 1 UIM domain, example: prod
But in that same UIM domain you can install additional hubs where needed, all these hubs will connect to the same domain: prod.
"If" you want to implement failover/HA, you follow the rules/doc you can find on these communities, but there you will still install your failover hub under/in the same domain and with another hubname.
There is a difference between the failover hub that will take over the tasks of the primary hub in case of failure and a normal hub that is only a collector for alarms and qos data from robots under these additional hub.

[MADANRAJ SIVAGNANAM]

But , we are configuring he robot.cfg like this in server after installing CA UIM agent in that server . we have both the hubs running simultaneously and if I give in the first line domain = NPCI_domain and hub = CHECAUIMHUB01_Pri_hub  the robot will report to primary and if i mention like below one
secondary_domain = HYDCLDUIMPHUB01_Domain ,secondary_hub = HYDCLDUIMPHUB01_hub the robor reports to secondary hub . This is how we are doing . In this case according to u where we have to use snmpcollector or snmp related probes ??




<controller>
domain = NPCI_domain
robotip =
hubip = 10.1.112.134
first_probe_port = 48000
hub = CHECAUIMHUB01_Pri_hub
hubrobotname = CHECAUIMHUB01
hubport = 48002
access_0 = 0
access_1 = 1
access_2 = 2
access_3 = 3
access_4 = 4
ip_version = ipv4
local_ip_validation = no
strict_ip_binding = no
robotname = HYDBSATMSW02
secondary_domain = HYDCLDUIMPHUB01_Domain
secondary_hub = HYDCLDUIMPHUB01_hub
secondary_hubrobotname = HYDCLDUIMPHUB01
secondary_hubip = 192.168.63.239
secondary_hubport = 48002
robotip_alias = 10.219.132.16
secondary_robotip_alias = 10.219.132.16
loglevel = 0
logsize = 1024
autoremove = no
suspend_on_loopback_only = yes
config_locking = no
hub_update_interval = 900
temporary_hub_broadcast = yes
system_uptime_qos = no
set_qos_source = no
unmanaged_security = domain_locked
alarm_level_comfail_restart =
alarm_level_dispatch_error = major
alarm_level_max_restarts = major
alarm_level_start_error = major
alarm_level_timed_not_finished = warning
alarm_level_unregister = major
alarm_level_request_error = major
alarm_level_timed_error_return = warning
alarm_level_postinstall = no alarm
robot_mode = normal
</controller>


Regards
Amar

[Luc Christiaens]

If this is the way you are working, and if the goal is that both hubs have a running data_engine and udm_manager AND if both hubs are pointing to the same SQL server; than you are working in the wrong way and you will probably need some consultancy to correct your environment.