[ZBX-15485] snmp responses randomly not processed Created: 2019 Jan 22  Updated: 2019 Jan 23  Resolved: 2019 Jan 23

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Server (S)
Affects Version/s: 4.0.2
Fix Version/s: None

Type: Problem report Priority: Trivial
Reporter: Dustin Howard Assignee: Unassigned
Resolution: Cannot Reproduce Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Zabbix 4.0 on CentOS Linux release 7.5.1804. Running VMware


Attachments: PNG File tcpdump.png     PNG File value-history.png    

 Description   

Only noticed with one paticular host that snmp responses are not processed to the point they are entered into the database. I noticed the issue since the graphs had "spotty" lines or information missing for random times. I am not using bulk requests. I see the packet coming into zabbix via tcpdump. I do not have debug info present on this ticket but I did increase debug level and compared all messages related to failed check with another check to the same host and seen no differences.

Checks are every 5 minutes. You can see in attached file "tcpdump.png" that packets are sent both ways for snmp check at 16:10 but no values exists in history. Graph history values in file attached "value-history.png"

I've increase debug levels but found nothing. I have also removed host and re-configured. Used bulk request and no Bulk requests. Can anyone point me to the next step?



 Comments   
Comment by Edgar Akhmetshin [ 2019 Jan 22 ]

Hello Dustin,

What version of the SNMP protocol is used? Also it would be great if you can provide a log file with a debug level of 4 and a traffic dump.

Regards,
Edgar

Comment by Oleksii Zagorskyi [ 2019 Jan 22 ]

2 items (with the same OID) are on the pictures - it misleads.
Looks like the 32-bit counter grows quite fast, which easily leads to the counter overflow, which, respectively, leads to one value skipping by zabbix. It's documented.

Comment by Dustin Howard [ 2019 Jan 22 ]

Thanks for your responses. I changed the item to pull the 64-bit counter instead and the issue has cleared.

Generated at Thu Apr 25 13:14:55 EEST 2024 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.