Same Value for any item - SNMPv1 OIDs

[knez]

Hi,

I made a template for some power inverters with SNMP v1. I created a few hosts too. But, the latest value shows the same value for all items... I have other templates with other hosts and all of them work just fine.

Here is a screenshot:

and after 30 sec (update time) it looks like this:



The expected value for the first item is 50Hz and the second item is 230 or 231 Volts (first is gathered in tenth of Hz so it has multiplier '0.1').

If someone has an idea or experience with these, that would be very useful for me.

Item looks like this:


and:

[1berto]

Post your item definitions...

[knez]

Post your item definitions...

I edited post... there they are... thanks...

[Markku]

What you are saying is that both OIDs return the same value, right? You said 30s later but the timestamps show something else. Can you post for example last 20 timestamps and values for both? (Latest data - Graph (for an item) - Values)

What does the MIB file say that the OIDs return?

Markku

[knez]

What you are saying is that both OIDs return the same value, right?

- Yes

Here are the values from the last 18 timestamp:

Timestamp	Phase AC Out Freq (multiplied by 0.1)	Phase V out
18.1.2020 13:18	50	500
18.1.2020 13:18	23	230
18.1.2020 13:17	50	500
18.1.2020 13:17	23	230
18.1.2020 13:16	23	230
18.1.2020 13:16	23	230
18.1.2020 13:15	23	230
18.1.2020 13:15	49	499
18.1.2020 13:14	50	500
18.1.2020 13:14	49	499
18.1.2020 13:13	23	231
18.1.2020 13:13	49	499
18.1.2020 13:12	23	230
18.1.2020 13:12	49	499
18.1.2020 13:11	49	499
18.1.2020 13:11	49	499
18.1.2020 13:10	49	499
18.1.2020 13:10	23	230

OID's:

OID	Object	Type	Values read by app OiDViEWFREE and expected too	Decription	Units
1.3.6.1.4.1.12551.4.1.2.2.1.6.1	tsiPhaseVout	INTEGER	230	Frequency of the output of the phase	0.1 Hz
1.3.6.1.4.1.12551.4.1.2.2.1.8.1	tsiPhaseACOutFreq	INTEGER	500	Output voltage of the phase	V

[Markku]

Ok, so each OID returns randomly the voltage or the frequency value, that seems very odd. And the documentation is off too: tsiPhaseVout has description of frequency, and vice versa. (Kind of logical: everything is messed up )

I would first make sure that I request only one of those values, for example just the voltage, from the device, and see if that makes any difference (= if the random values continue or not). And I would try to repeat this random behaviour with simple snmpwalk commands so that the product vendor could be contacted with a bug report. Firmware upgrade for the device could make a difference.

Markku

[knez]

And the documentation is off too: tsiPhaseVout has description of frequency, and vice versa. (Kind of logical: everything is messed up

this is my mistake.. sorry.. it is a correct description in OID table:

I tried with snmpget and snmpwalk and values are always correct and expected. This app "OiDViEWFREE" show values correct too.

Milan

[Markku]

Especially in that situation, please delete or disable polling of the frequency value and confirm that when Zabbix is only polling the voltage the values are correct. Then do the same other way around = disable voltage and poll only frequency values, check that the values are fine.

I haven't ever seen Zabbix intermixing values between polled items. That said, my SNMP items have always been SNMP v2, if it makes a difference.

Btw in the screenshot the tsiPhaseVout units are said to be 0.1 V, meaning that raw number 230 would mean 23.0 V...

Markku

[Markku]

MIB can be found in: http://www.acbackuptsi.com/doc/MIB/CET-TSI-MIB.mib

It says:

        --OID : .1.3.6.1.4.1.12551.4.1.2.2.1.6
        tsiPhaseVout OBJECT-TYPE
      SYNTAX                                            Integer32
      UNITS                                                "0.1V"
      MAX-ACCESS                                    read-only
      STATUS                                            current
      DESCRIPTION                                    "Output voltage of the phase"

        --OID : .1.3.6.1.4.1.12551.4.1.2.2.1.8
        tsiPhaseACOutFreq   OBJECT-TYPE
      SYNTAX                                            Integer32
      UNITS                                                "0.1Hz"
      MAX-ACCESS                                    read-only
      STATUS                                            current
      DESCRIPTION                                    "Frequency of the output of the phase"

Have you tried in snmpwalk and in Zabbix without the .1 at the end of the OID in the item configurations? I'm not totally sure about the exact syntax there.

Markku

[knez]

That "0.1Hz" is a vendor mistake ... it is just "Hz", actually.

When OID 1 is disabled and OID 2 enabled and vice versa I got correct values ...

[knez]

Have you tried in snmpwalk and in Zabbix without .1?

I did:

[Markku]

Interesting. How about disabling bulk requests in the host SNMP interface? Bulk requests should not be available in SNMP v1, but it's worth trying.

Did you already try SNMP v2c? Please do if the device supports it.

And definitely, now it's a good time to run some tcpdump or wireshark/tshark to capture the actual SNMP traffic between Zabbix server and the affected device, when you have both items enabled and you experience the incorrect values. Something like "tcpdump -s 0 -w capturefile.pcap host 192.168.132.253 and port 161", run 5-10 minutes or until you see the incorrect values in Zabbix data. Then use Wirehark to see the actual contents of the SNMP requests and responses in the capture file. Do not post the capture file here, unless you can first change the SNMP community to something temporary and you are sure that the capture does not contain anything sensitive.

Markku

[knez]

Interesting. How about disabling bulk requests in the host SNMP interface?

Markku,

this is it. I'm new in SNMP and in Zabbix too and didn't know about bulk requests. Thank you very much for your help.

I'll try to capture SNMP traffic when I have a little more experience with linux, SNMP and Zabbix...