Hi,
In my NetXMS (v1.2.8 ) application I would like to detect the SNMP drop out events.
I have tried to use the data collection error feature from Data Collection Configuration/ Edit Threshold but not generated any error event.
The applied condition was: Function: Data Collection error; Samples: 1; Value: >0; activation event: SNMP_LINK_DOWN or SYS_SNMP_UNREACHABLE.
What is the correct configuration for this type of error event?
Thank you in advance and best regards.
Hi!
If I understand you correctly, you are trying to process SNMP traps, right? If this is the case, you don't need data collection - you have to configure trap mapping and process events accordingly. System already have predefined mappings for standard SNMP traps like "interface down". Check SNMP trap log to ensure that server receives SNMP traps.
Best regards,
Victor
Hi,
Thank you for your fast reply and sorry for the misunderstanding. I would like to controll some equipment by SNMP using DCI for key parameters. Unfortunately sometimes it happens that the controlled device not responding to SNMP GET request generated by DCI. In this case the IP connection is alive but the SNMP server integrated in controlled equipment is in frozen state (have to use power reset to restart the SNMP server from controlled equipment). So, I would like to detect this error condition. For this I tried to use the data collection error feature from Data Collection Configuration/ Edit Threshold but not generated any alarm event. Very interesting that during this error state in the last values window is appear the error notification but in alarm list can not be found any alarm notification referring to this SNMP request error.
Best regards,
Zsolt
Hi!
Looks like a bug - maybe error threshold is broken somehow. I'll try to reproduce this on my test system.
Also, you should get SYS_SNMP_DOWN event at next configuration poll if SNMP agent is not responding - do you get those events?
Best regards,
Victor
Hi,
During under the SNMP problem I made a detailed investigation with the following results:
In the Last Values window I found that in case of all DCI the content of "Value" column is <<ERROR>> (It's OK), the content of "Threshold" column in each DCI are OK (it's normal because no treshold violation) except the DCI what is configured with Data Collection Error threshold. In this case the "Threshold" column contain the notification of error state: ! error(1)>0 (it's OK)
The alarm window of problematic node is empty despite the fact that it is an outstanding alarm. Also in case of problematic node from Objects window I do not see any graphical notification about the pending alarm (these are the main problems).
In Event Log window I found the SYS_SNMP_UNREACHABLE event without SYS_SNMP_OK. These activation events had set in the data collection error DCI.
Thank you in advance and best regards,
Zsolt
I suspect that your setup is not entirely correct. First of all, it is not recommended to use system events in data collection thresholds - better create your own event. And second, event generation does not automatically cause alarm - you have to create alarm from event in event processing policy.
Best regards,
Victor
Hi,
Thanks for your help. I have created an own event for this error condition and I configured properly. Then the alarm event generation is works fine.
Best regards,
Zsolt