Bug ID 707675: FQDN nodes or pool members flap when DNS response received

Last Modified: May 14, 2019

Bug Tracker

Affected Product:  See more info
BIG-IP LTM(all modules)

Known Affected Versions:
12.1.0, 12.1.0 HF1, 12.1.0 HF2, 12.1.1, 12.1.1 HF1, 12.1.1 HF2, 12.1.2, 12.1.2 HF1, 12.1.2 HF2, 12.1.3, 12.1.3.1, 12.1.3.2

Fixed In:
12.1.3.3

Opened: Feb 23, 2018
Severity: 3-Major

Symptoms

When an LTM pool is configured with FQDN nodes or pool members, the LTM pool and associated virtual server(s) may transition from an UP to DOWN state and back over a period of a few seconds. Such an event is accompanied by log messages similar to the following: -- notice mcpd[#]: 01071682:5: SNMP_TRAP: Virtual /Common/vs_test has become unavailable -- notice mcpd[#]: 010719e7:5: Virtual Address /Common/123.45.67.89 general status changed from GREEN to RED. -- notice mcpd[#]: 010719e8:5: Virtual Address /Common/123.45.67.89 monitor status changed from UP to DOWN. -- err mcpd[#]: 01020066:3: The requested Pool Member (/Common/Test_Pool /Common/test-dummy.com-12.34.56.78 443) already exists in partition Common. -- notice bigd[##]: 01060144:5: Pool /Common/Test_Pool member /Common/test-dummy.com-12.34.56.78 session status enabled by monitor -- notice bigd[##]: 01060145:5: Pool /Common/Test_Pool member /Common/test-dummy.com-12.34.56.78 monitor status up. [ /Common/mon_test_https: UP ] [ was checking for 0hr:0min:2sec ] -- notice mcpd[#]: 01071681:5: SNMP_TRAP: Virtual /Common/vs_test has become available -- notice mcpd[#]: 010719e7:5: Virtual Address /Common/123.45.67.89 general status changed from RED to GREEN. -- notice mcpd[#]: 010719e8:5: Virtual Address /Common/123.45.67.89 monitor status changed from DOWN to UP. This symptom repeats each time a DNS query is performed to resolve the FQDN node/pool-member name to its IP addresses, based on the 'interval' value configured for the FQDN node. This symptom occurs only when the 'autopopulate' value is set to 'enabled' for the FQDN node/pool-member.

Impact

LTM pool and virtual server are briefly and periodically marked DOWN. Traffic may be impacted.

Conditions

-- LTM pool is configured with FQDN nodes or pool members. -- The 'autopopulate' value is set to 'enabled' for the FQDN node/pool-member.

Workaround

Either of the following methods can be used to work around this issue: -- Configure static IP addresses instead of FQDN nodes/pool-members. -- Set the 'autopopulate' value to 'disabled' for the FQDN node/pool-member, if possible (that is, if only one IP address is required/expected to be returned for the FQDN name, which means that the 'autopopulate' feature of FQDN nodes/pool-members is not required).

Fix Information

FQDN node/pool-member and corresponding pool and virtual server are no longer briefly marked DOWN when the DNS server is queried to resolve the FQDN name, with the 'autopopulate' feature enabled for the FQDN node/pool-member. This issue is resolved by the FQDNv2 feature re-implementation in this version of the software.

Behavior Change