This article describes the information on why SNMPD is causing high RE CPU utilisation on SRX.

When snmp v3-informs are configured on SRX without a trap destination for a trap source then this can cause snmpd to consume a high CPU which can even reach 100%. However this issue is not seen with SNMP non-v3-informs configuration.

> SNMP config which triggers the issue
set snmp v3 notify SNMP-V2-INFORM-NOTIFY type inform

>> snmpd daemon on the SRX during the issue: 
user@Device> show system processes extensive
last pid:  2748;  load averages:  1.46,  1.31,  1.29  up 7+11:03:10    14:04:37
372 processes: 6 running, 338 sleeping, 28 waiting

Mem: 268M Active, 4464M Inact, 960M Wired, 151M Buf, 26G Free
Swap: 8192M Total, 8192M Free
  PID USERNAME    PRI NICE   SIZE    RES STATE   C   TIME    WCPU COMMAND
9564 root        103    0   744M 33620K CPU3    3 135.9H 100.00% snmpd    <<<<<<<<<<<<<<<<
   11 root        155 ki31     0K    64K RUN     1 134.2H  74.37% idle{idle: cpu1}
 

> High RE cpu on the SRX

This issue is being tracked in PR 1691986.
 

Workaround:

Use traps instead of SNMP v3 informs in the following config:
"set snmp v3 notify SNMP-V2-INFORM-NOTIFY type trap"


Resolved in: 

junos:20.4R3-S8
junos:21.2R3-S6
junos:22.4R3
junos:23.1R2
junos:23.2R1
 

2023-12-18 : Article Created
2024-01-02: Article published
2026-01-29 : Updated Category