Cisco Bug: CSCvu04665 - CUBE keeps sending REINVITES to peer legs leading to high CPU and eventually crashes.
Aug 29, 2020
- Cisco IOS
- Cisco Catalyst 9400 Supervisor Engine-1XL-Y
- Cisco Catalyst 9300-48U-A Switch
- Cisco Catalyst 9300-48UN-A Switch
- Cisco Catalyst 9300-48UXM-A Switch
- Cisco Catalyst 9300-48P-A Switch
- Cisco Catalyst 9300-48T-A Switch
- Cisco Catalyst C9500-16X-E Switch
- Cisco Embedded Wireless Controller on Catalyst 9120AX Access Points
- Cisco Catalyst 9130AXE Access Point
Known Affected Releases
Symptom: 1. Multiple REINVITES triggered by CUBE in a loop towards both peer legs. Each 200OK response from peer-leg leads to the generation of a REINVITE on the other leg. Sample call-flow: IP phone -- sip -- CUCM -- sip -- CUBE -- sip -- server-group -- ITSP SBC 1 and ITSP SBC2 2. High CPU in CCSIP_SPI_CONTOL process: qwerty#show processes cpu sorted | ex 0.0 CPU utilization for five seconds: 94%/0%; one minute: 33%; five minutes: 13% PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process 732 162116 48235 3360 66.10% 22.00% 5.29% 0 CCSIP_SPI_CONTRO 734 14420 38068 378 1.68% 0.59% 0.13% 0 CCSIP_UDP_SOCKET Will also see an abnormal values for message rate but CPS will not increase paralelly. Outputs to check: show sip-ua history stats message-rate table show call history stats cps table Router can just crash as well depending on the original call volume. Conditions: This scenario is hit only when the following conditions are true: 1. Server groups are configured on outgoing dial-peer. 2. Element 1 (chosen by server-group) sends a 18X w/sdp followed by 4xx/5xx msg. Without a 18x msg with SDP received from element 1, CUBE will not trigger REINVITES post call-connect.
Bug details contain sensitive information and therefore require a Cisco.com account to be viewed.
Bug Details Include
- Full Description (including symptoms, conditions and workarounds)
- Known Fixed Releases
- Related Community Discussions
- Number of Related Support Cases