Re: [Freeipmi-users] Interpretation of IPMI events for monitoring
Hello Al, Many thanks for your quick feedback. I will forward to monitoring software vendor. Mike -Original Message- From: Al Chu Sent: Wednesday, July 14, 2021 7:55 PM To: FRANK Michael ; freeipmi-users@gnu.org Subject: Re: [Freeipmi-users] Interpretation of IPMI events for monitoring Hi Frank, > Is it correct that the meaning of "transition to" is, that the sensor > was before in a different state and now has the reported state? This is my interpretation of the specification. > Current situation is that the monitoring software interpret the event > "transition to OK" as an CRITCAL state which I think is not correct. With the caveat that every manufacturer / sensor could have alternate interpretations, I've always gone with the belief that "transition to OK" is "nominal" / "ok" as a monitoring state. It's what I set as the default within the libipmimonitoring interpretations (i.e. ipmi-sensors --output-sensor-state). Al On Wed, 2021-07-14 at 15:13 +, FRANK Michael wrote: > Hello, > > I am currently in the discussion with monitoring software developers > about the interpretation of sensor events. > In general what I understood is: an event only changes in the > situation the state of a sensor changes but the state of the sensor > is not "really" checked each time we read the sensor e.g with ipmi- > sensor. My abstract imagination is that the BMC has a table of > sensorĀ“s and the latest events and if the state of a sensor changes > the BMC update the according entry in this table. > > Main discussion is about the event "transition to..." and especially > the event "transition to OK" (Type code 07h, Offset 00h). > Unfortunately, there are no further explanations in the IPMI Specs > Sensor and Event Code Tables. > > So here is my question: > Is it correct that the meaning of "transition to" is, that the sensor > was before in a different state and now has the reported state? > For example, event "transition to OK" means the sensor had, before > the event changed, a state other than OK, like Critical, Non-Critical > etc. and his present state is now OK. > > I would much appreciate if an expert could verify the above > statement. > Current situation is that the monitoring software interpret the event > "transition to OK" as an CRITCAL state which I think is not correct. > > ID | Name | Type | Reading| Units | Event > 80 | PS2 12V UV Fault | Power Supply | N/A| N/A | > 'transition to OK' > > Record ID: 80 > Record Type: Compact Sensor Record (2h) > ID String: PS2 12V UV Fault > Sensor Type: Power Supply (8h) > Sensor Number: 122 > IPMB Slave Address: 10h > Sensor Owner ID: 20h > Sensor Owner LUN: 0h > Channel Number: 0h > Entity ID: power supply (10) > Entity Instance: 2 > Entity Instance Type: Physical Entity > Event/Reading Type Code: 7h > Sensor Direction: Unspecified > Assertion Event Enabled: 'transition to OK' > Assertion Event Enabled: 'transition to Non-Critical from OK' > Assertion Event Enabled: 'transition to Critical from less severe' > Assertion Event Enabled: 'transition to Critical from Non- > recoverable' > Assertion Event Enabled: 'Monitor' > Deassertion Event Enabled: 'transition to OK' > Deassertion Event Enabled: 'transition to Non-Critical from OK' > Deassertion Event Enabled: 'transition to Critical from less severe' > Deassertion Event Enabled: 'transition to Critical from Non- > recoverable' > Deassertion Event Enabled: 'Monitor' > Share Count: 1 > ID String Instance Modifier Type: Numeric > ID String Instance Modifier Offset: 0 > Entity Instance Sharing: Same for all records > Sensor Event: 'transition to OK' > > 10.240.4.12: = > 10.240.4.12: IPMI 1.5 Get Channel Authentication Capabilities Request > 10.240.4.12: = > 10.240.4.12: RMCP Header: > 10.240.4.12: > 10.240.4.12: [ 6h] = version[ 8b] > 10.240.4.12: [ 0h] = reserved[ 8b] > 10.240.4.12: [ FFh] = sequence_number[ 8b] > 10.240.4.12: [ 7h] = message_class.class[ 5b] > 10.240.4.12: [ 0h] = message_class.reserved[ 2b] > 10.240.4.12: [ 0h] = message_class.ack[ 1b] > 10.240.4.12: IPMI Session Header: > 10.240.4.12: > 10.240.4.12: [ 0h] = authentication_type[ 8b] > 10.240.4.12: [ 0h] = session_sequence_number[32b] > 10.240.4.12: [ 0h] = session_id[32b] > 10.240.4.12: [ 9h] = ipmi_msg_len[ 8b] > 10.240.4.12: IPMI Message Header:
Re: [Freeipmi-users] Interpretation of IPMI events for monitoring
Hi Frank, > Is it correct that the meaning of "transition to" is, that the sensor > was before in a different state and now has the reported state? This is my interpretation of the specification. > Current situation is that the monitoring software interpret the event > "transition to OK" as an CRITCAL state which I think is not correct. With the caveat that every manufacturer / sensor could have alternate interpretations, I've always gone with the belief that "transition to OK" is "nominal" / "ok" as a monitoring state. It's what I set as the default within the libipmimonitoring interpretations (i.e. ipmi-sensors --output-sensor-state). Al On Wed, 2021-07-14 at 15:13 +, FRANK Michael wrote: > Hello, > > I am currently in the discussion with monitoring software developers > about the interpretation of sensor events. > In general what I understood is: an event only changes in the > situation the state of a sensor changes but the state of the sensor > is not "really" checked each time we read the sensor e.g with ipmi- > sensor. My abstract imagination is that the BMC has a table of > sensorĀ“s and the latest events and if the state of a sensor changes > the BMC update the according entry in this table. > > Main discussion is about the event "transition to..." and especially > the event "transition to OK" (Type code 07h, Offset 00h). > Unfortunately, there are no further explanations in the IPMI Specs > Sensor and Event Code Tables. > > So here is my question: > Is it correct that the meaning of "transition to" is, that the sensor > was before in a different state and now has the reported state? > For example, event "transition to OK" means the sensor had, before > the event changed, a state other than OK, like Critical, Non-Critical > etc. and his present state is now OK. > > I would much appreciate if an expert could verify the above > statement. > Current situation is that the monitoring software interpret the event > "transition to OK" as an CRITCAL state which I think is not correct. > > ID | Name | Type | Reading| Units | Event > 80 | PS2 12V UV Fault | Power Supply | N/A| N/A | > 'transition to OK' > > Record ID: 80 > Record Type: Compact Sensor Record (2h) > ID String: PS2 12V UV Fault > Sensor Type: Power Supply (8h) > Sensor Number: 122 > IPMB Slave Address: 10h > Sensor Owner ID: 20h > Sensor Owner LUN: 0h > Channel Number: 0h > Entity ID: power supply (10) > Entity Instance: 2 > Entity Instance Type: Physical Entity > Event/Reading Type Code: 7h > Sensor Direction: Unspecified > Assertion Event Enabled: 'transition to OK' > Assertion Event Enabled: 'transition to Non-Critical from OK' > Assertion Event Enabled: 'transition to Critical from less severe' > Assertion Event Enabled: 'transition to Critical from Non- > recoverable' > Assertion Event Enabled: 'Monitor' > Deassertion Event Enabled: 'transition to OK' > Deassertion Event Enabled: 'transition to Non-Critical from OK' > Deassertion Event Enabled: 'transition to Critical from less severe' > Deassertion Event Enabled: 'transition to Critical from Non- > recoverable' > Deassertion Event Enabled: 'Monitor' > Share Count: 1 > ID String Instance Modifier Type: Numeric > ID String Instance Modifier Offset: 0 > Entity Instance Sharing: Same for all records > Sensor Event: 'transition to OK' > > 10.240.4.12: = > 10.240.4.12: IPMI 1.5 Get Channel Authentication Capabilities Request > 10.240.4.12: = > 10.240.4.12: RMCP Header: > 10.240.4.12: > 10.240.4.12: [ 6h] = version[ 8b] > 10.240.4.12: [ 0h] = reserved[ 8b] > 10.240.4.12: [ FFh] = sequence_number[ 8b] > 10.240.4.12: [ 7h] = message_class.class[ 5b] > 10.240.4.12: [ 0h] = message_class.reserved[ 2b] > 10.240.4.12: [ 0h] = message_class.ack[ 1b] > 10.240.4.12: IPMI Session Header: > 10.240.4.12: > 10.240.4.12: [ 0h] = authentication_type[ 8b] > 10.240.4.12: [ 0h] = session_sequence_number[32b] > 10.240.4.12: [ 0h] = session_id[32b] > 10.240.4.12: [ 9h] = ipmi_msg_len[ 8b] > 10.240.4.12: IPMI Message Header: > 10.240.4.12: > 10.240.4.12: [ 20h] = rs_addr[ 8b] > 10.240.4.12: [ 0h] = rs_lun[ 2b] > 10.240.4.12: [ 6h] = net_fn[ 6b] > 10.240.4.12: [ C8h] = checksum1[ 8b] > 10.240.4.12: [ 81h] = rq_addr[ 8b] > 10.240.4.12: [ 0h] = rq_lun[ 2b] > 10.240.4.12: [ 31h] = rq_seq[ 6b] > 10.240.4.12: IPMI Command Data: > 10.240.4.12: -- > 10.240.4.12: [ 38h] = cmd[ 8b] > 10.240.4.12: [ Eh] = channel_number[ 4b] > 10.240.4.12: [ 0h] = reserved1[ 3b] > 10.240.4.12: [ 1h] = get_ipmi_v2.0_extended_data[ 1b] >