Unsupported items due to "Value has unknown type 0x05"

[mmussons]

Hi,
sometimes (this is not the first) we get without any apparent reason a lot of unsupported items from a lot of different hosts.

For example in latest data we have values:

Timestamp Value
03/05/2020 04:25:33 AM NORMAL (0)
03/05/2020 03:25:33 AM NORMAL (0)
03/05/2020 02:25:33 AM NORMAL (0)
03/05/2020 01:25:32 AM NORMAL (0)

and suddenly in the next polling time for the same host:

12120:20200305:052529.267 error reason for "032489_01:CFG_SYNC" changed: Value has unknown type 0x05
12120:20200305:052529.267 item "032489_01:EDO_ALM_METER6" became not supported: Value has unknown type 0x05

I can share with you a lot of messages like this:

12120:20200305:052535.538 item "083671_01:SPV_PLC_TOUTDOOR" became not supported: Value has unknown type 0x05
12120:20200305:052535.538 item "083671_01:EAO_PLC_ZONE" became not supported: Value has unknown type 0x05
12120:20200305:052535.538 item "083671_01:EDO_PLC_BLOCKED" became not supported: Value has unknown type 0x05
12120:20200305:052535.538 item "083671_01:EDO_PLC_COMFORT" became not supported: Value has unknown type 0x05
12120:20200305:052535.538 item "083671_01:SDO_PLC_COMFORT" became not supported: Value has unknown type 0x05
12120:20200305:052535.538 item "083671_01:SDO_PLC_RESET" became not supported: Value has unknown type 0x05
12120:20200305:052535.577 item "083671_01:EDO_CHATER1" became not supported: Value has unknown type 0x05
12118:20200305:052536.472 error reason for "083677_01:CFG_SYNC" changed: Value has unknown type 0x05
12118:20200305:052536.472 item "083677_01:EAO_ALM_PRGINDEX" became not supported: Value has unknown type 0x05
12118:20200305:052536.472 item "083677_01:EAO_ALM_PRGTYPE" became not supported: Value has unknown type 0x05
12118:20200305:052536.472 item "083677_01:EAO_GPRS_LEVEL" became not supported: Value has unknown type 0x05
12118:20200305:052536.472 item "083677_01:EAO_METER18" became not supported: Value has unknown type 0x05

We have thousends of hosts maganed most of them using snmp protocol
We have Zabbix 3.0.5 and search for a fix or any information in Zabbix doc but not comments.

Any idea?
Thanks in advance.

[dimir]

That means snmp agent returned error or exception. How often does this happen? Can you reproduce it with snmpget/snmpwalk to get more information on the error?

[mmussons]

I can see these message from yesterday morning over 06:13AM until today at 05:25AM more or less, but not before this March.
I have counted them:

65 20200304:06
1 20200304:07
134 20200304:11
62 20200305:00
49 20200305:01
297 20200305:05

I can't reproduce because when I request values manually always goes good.

Maybe is due to a high volume of requests?
Can I change this behaviour using config file?

[dimir]

You could try turning "Bulk requests" off if you have it enabled: https://www.zabbix.com/documentation...ulk_processing

[mmussons]

No, it's impossible.
We are getting values of more than 5000 hosts with templates with more than 500 items using snmp (more than half of them every hour); without bulk requests in the past we had firewall problems and a lot of network traffic. And also we will add 1500 hosts more in a short future...
Your proposal may be worse than the current problem.

Do you think is a snmp agent error?
It's possible to configure some retrys like "Unreachable/unavailable" hosts?

[dimir]

I'm sure this is an error coming from snmp agent. In your case I'd go through all the frequencies of your snmp checks, are they not too frequent? There's no way to configure SNMP retries unfortunately. The timeout is whatever is configured on your server/proxy.

[mmussons]

I will follow the problem during one or two weeks and then depending on the frequency I will take a decision, maybe increase polling time or maybe change bulk request but I will have to do a lot of tests in a many points in the network.
Thank you.