SNMP error: (genError) A general failure occured

[freeco]

Hi,

I have a error on my fresh install of zabbix. In a VM, I installed a Debian 8.8 from the network boot cd.

Then i put the repo of Debian and i installed Zabbix from the official repository of zabbix.com.

I imported template from Zabbix share (https://share.zabbix.com/network-app...-snmp-template). The discover work fine and after i've got this error : SNMP error: (genError) A general failure occured

I try this OID with snmpwalk command line and i've result.

Screenshot of my problem :


Have you an idea ?


Debug log on zabbix server :

 30406:20170530:121003.156 End of zbx_vc_add_value():FAIL
 30391:20170530:121003.156 In substitute_simple_macros() data:'{$SNMP_COMMUNITY}'
 30390:20170530:121003.156 In zbx_snmp_get_octet_string()
 30389:20170530:121003.156 End of zbx_snmp_close_session()
 30392:20170530:121003.156 End of zbx_snmp_translate() oid_translated:'1.3.6.1.4.1.16139.2.23.1.5.1.2'
 30406:20170530:121003.156 In zbx_vc_add_value() itemid:25656 value_type:1 timestamp:1496139003.107724694
 30391:20170530:121003.156 In DCget_user_macro() macro:'{$SNMP_COMMUNITY}'
 30390:20170530:121003.156 zbx_snmp_get_octet_string() full value:'STRING: "S1-VMFS12-T"' hint:'(null)'
 30389:20170530:121003.156 getting SNMP values failed: SNMP error: (genError) A general failure occured

[freeco]

I've found solution. I disabled "Use bulk requests" on host, and it worked.

[WytheNet]

Thank you！

[c1support]

I've found solution. I disabled "Use bulk requests" on host, and it worked.

Thanks a lot.
I had a similar (the same?) issue with Zabbix v5.0.9 and I can confirm that disabling "Use bulk requests" on the affected host (under "Configuration" => "Hosts" => [affected host] => "SNMP" => uncheck "use bulk requests") worked.



Edit:
Unfortunately, after unchecking "Use bulk requests" on all my SNMP-devices resulted in higher loads for Zabbix Server:
- "Utilization of poller data collector processes" immediately raised from 20% to 95%
Adding more pollers in /etc/zabbix/zabbix_server.conf (from 10 to 50) only reduced this value to 80%

- "CPU load average (1m)" immediateo raised from 0.8 to 7, and after about 2 hours to 30 (even though Zabbix was untouched in these 2 hours)
Adding more pollers did not affect this value.


This lead me to the conculusion that I only uncheck "Use bulk requests" for initiating a new host. This way, all the values will be properly discovered by Zabbix Server.
After this is completed, "Use bulk requests" will be checked again.


There seems to be no entry related to this behaviour (admittedly v5.0.9 is dating back some 18 months - will try again after the next update):

https://www.zabbix.com/documentation/5.0/en/manual/installation/known_issues

=> If anybody has a more elegant way, hints will be appreciated.

[c1support]

For future reference, I wanted to add details on the system:
- 4 Cores, 4GB RAM
- Zabbix v5.0.9
- 40 Hosts, 15'000 Items
- Value cache effectiveness: Zabbix value cache hits: 130 vps
- Values processed per second: values processed per second: 30
- Required server performance, new values per second: 225

Most of items are collected via SNMPv3, and most of the SNMPv3-values are gathered from the devices where the feature "use bulk requests" was temporatly modified.

The official Zabbix documentation details requirements (small, medium, large, very large systems):

https://www.zabbix.com/documentation/5.0/en/manual/installation/requirements

this is considered a "small" system (even though the only parameter taken into consideration is # of Monitored Hosts - # NVPS is ignored).

In an old forum entry, this blogpost article was referenced:
https://blog.zabbix.com/scalable-zab...400-nvps/2615/
As early as 2013, Zabbix was able to cover ~10'000 NVPS.


=> If anybody can see an obviuos error I performed, please enlighten us. Thanks a lot!

[dimir]

There might come some improvements on bulk SNMP monitoring, you are welcome to follow the discussion here:

https://support.zabbix.com/browse/ZBXNEXT-4428