[ZBX-15391] SNMPv3 Checks report "Timeout while connecting to "host:port"." Created: 2019 Jan 02  Updated: 2019 Jan 05  Resolved: 2019 Jan 02

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Server (S)
Affects Version/s: 4.0.3
Fix Version/s: None

Type: Problem report Priority: Minor
Reporter: Igor Assignee: Unassigned
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Ubuntu 18.04.1 LTC Bionic

Attachments: PNG File monitored_host.PNG     PNG File zabbix_server.PNG     PNG File zabbix_server_working.PNG    
Issue Links:
Duplicate
duplicates ZBX-2152 multiple SNMPv3 checks get unexpected... Closed

 Description   

Problem:
We are currently monitoring 30 hosts using a custom built SNMPv3 template with 4 items. On about 12 of those hosts we get the error message "Timeout while connecting to "host:port".".

Troubleshooting done:
The hosts that are working and that aren't working are using completely the same templates with no modifications done to specific hosts.

We've recreated hosts that aren't working by cloning them from hosts that are.

We've tried changing UDP ports that the hosts are listening on (and then making the corresponding Port change for an SNMP interface for a host in Zabbix, we did not modify the template Port for this test, just the host port).

We've also captured network traffic between the Zabbix server and the hosts that are and aren't working properly.

While capturing traffic on the hosts that didn't work we could see traffic and information sent and received by both the Zabbix server and the host that was being monitored. However Zabbix was still reporting the above error.

We were able to notice a difference between the packets generated by the Zabbix server that were sent to the hosts that were and weren't working.

On the hosts that are working correctly the Zabbix server seemed to respect the "Use bulk requests" setting and it was sending the request all in one packet.

While the hosts that didn't work properly the request was broken up into individual packets even though "Use bulk requests" was checked. Turning off bulk requests didn't fix the issue either.
The individual requests were still being received by the monitored host though, as shown in screenshots bellow, and the responses were being sent back.

 Comments   
Comment by Igor [ 2019 Jan 02 ]

Forgot to mention that running snmpwalk from the zabbix server to any of the monitored hosts works without any problems as well.

Comment by Oleksii Zagorskyi [ 2019 Jan 02 ]

Bulk is not related here.

$ snmptranslate 1.3.6.1.6.3.15.1.1.4.0
SNMP-USER-BASED-SM-MIB::usmStatsUnknownEngineIDs.0

$ snmptranslate 1.3.6.1.6.3.15.1.1.2.0
SNMP-USER-BASED-SM-MIB::usmStatsNotInTimeWindows.0

usmStatsNotInTimeWindows most likely indicates that you have duplicated EngineIDs for those 12 hosts. See ZBX-2152
It's stated in documentation that all devices should use unique EngineID.

It's also possible that those devices are buggy, see ZBX-8385

Comment by Oleksii Zagorskyi [ 2019 Jan 02 ]

I'm pretty sure it's not a bug, so I close it.

Comment by Igor [ 2019 Jan 05 ]

Oh man has this been a revelation. I know that there is little you guys can do about it but I sure wish that the error message in Zabbix was different. I spent so much time chasing down this problem and going down the wrong path. This was it for me and I was able to fix the issue on my end now that I knew what I was looking for.

Thanks for finally putting me on the right path though.

Comment by richlv [ 2019 Jan 05 ]

Igor, you might want to vote on ZBX-8385.

Oleksiy, awesome job on SNMPv3 issues, as usual

Comment by Igor [ 2019 Jan 05 ]

Voted. I would like to vote 10 times for this if I could. It took me 2 hours to fix the issue once Oleksiy let me know what the problem actually was. I've been troubleshooting connectivity issues for a month before that.

Generated at Mon May 19 06:45:49 EEST 2025 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.