[ZBX-8191] Gaps in SNMP graphs after upgrade to 2.2.3 Created: 2014 May 10 Updated: 2017 May 30 Resolved: 2014 May 15 |
|
Status: | Closed |
Project: | ZABBIX BUGS AND ISSUES |
Component/s: | Server (S) |
Affects Version/s: | 2.2.3 |
Fix Version/s: | None |
Type: | Incident report | Priority: | Major |
Reporter: | Tomasz Grzywacz | Assignee: | Unassigned |
Resolution: | Duplicate | Votes: | 0 |
Labels: | bulk, snmp | ||
Remaining Estimate: | Not Specified | ||
Time Spent: | Not Specified | ||
Original Estimate: | Not Specified |
Attachments: |
![]() ![]() ![]() ![]() |
||||||||
Issue Links: |
|
Description |
When I upgraded from 2.2.2 to 2.2.3 I receive a lot of erros like: 3656:20140509:001126.314 temporarily disabling SNMP agent checks on host "PPPoE Krasow - Area9": host unavailable 3659:20140509:001054.749 SNMP agent item "nstremeDual" on host "PPPoE Krasow - Area9" failed: another network error, wait for 15 seconds 3666:20140509:002416.057 cannot send list of active checks to [127.0.0.1]: host [pies] not found but also : 3660:20140509:001121.957 enabling SNMP agent checks on host "PPPoE Dziezgow - Area10": host became available As result all SNMP hosts are being disabled, but from time to time become also enabled (flapping). Anyway graphs for related SNMP items contain holes or nothing is drawn. Issue is related with SNMP checks only, all items for zabbix agent based items are graphed properly. The configuration of zabbix server is identical, as in case of 2.2.2 (which is working fine). As a trial I changed binaries only to 2.2.2 version and there is no issue. For sure this is not an issue with SNMP connectivity ( I checked it on linux console with snmpwalk and snmpget) |
Comments |
Comment by Aleksandrs Saveljevs [ 2014 May 10 ] |
Tomasz, it would be nice if you could provide more low-level information. For instance, DebugLevel=4 and tcpdump of the traffic between Zabbix and one of these devices. |
Comment by Aleksandrs Saveljevs [ 2014 May 10 ] |
This issue seems to be from the same category as |
Comment by Marc [ 2014 May 10 ] |
asaveljevs, I noticed this as well for Brocade fiber channel switches beside high CPU utilization on NetApps. |
Comment by Tomasz Grzywacz [ 2014 May 11 ] |
I really cannot see anything special from debug log. |
Comment by Aleksandrs Saveljevs [ 2014 May 12 ] |
According to tcpdump, there are a lot of replies with "noSuchName" error status. Is my understanding correct that you are using SNMPv2 for monitoring this device? If so, what happens when you switch to SNMPv1? Are you using any SNMP proxy, as in |
Comment by Tomasz Grzywacz [ 2014 May 12 ] |
Aleksandrs, I post also a thread in forum related to my issue. https://www.zabbix.com/forum/showthread.php?p=149711#post149711 User aib just confirmed, that this is known issue in 2.2.3 Can anyone from Zabbix development team confirm this fact? |
Comment by Aleksandrs Saveljevs [ 2014 May 13 ] |
Well, we can only confirm that some devices out there do not conform to SNMP standard and that is why Zabbix is having an issue with them. In this particular case, your device, even though monitored through SNMPv2, seems to behave like SNMPv1, that is why I suggested to try SNMPv1. |
Comment by Aleksandrs Saveljevs [ 2014 May 15 ] |
In Zabbix 2.4, which is soon to be released, we shall try to provide a way to enable and disable SNMP bulk for each interface. Please follow progress in |