Symptoms

Per general recommendation, Pools should contain enough IP addresses to cover the number of interfaces added to the pool. For optimal load-balancing if a node fails, dynamic IP pools require N*(N-1) IP addresses, where N is the number of nodes or interfaces that participate in the pool. NOTE: The recommendation does not prohibit or restrict creating an IP pool with a single IP across multiple node interfaces. Such setups may be used for third-party solution integration. Recurring alerts similar to the following are seen when there is no issues with physical interfaces: Aggregate network link 10gige-agg-1 (lagg0) error, interface(s) down: all; no addresses Cluster-4# isi events show 1.7434 ID: 1.7434 Type: 200020004 Severity: critical Value: 2.0 Message: Aggregate network link 10gige-agg-1 (lagg0) error, interface(s)... Node: 1 Lifetime: Fri Dec 9 09:59:22 2016 - Now Quieted: Not quieted Specifiers: lnn: 1 thresh: 1.0 phys_iface: '' val: 2 nic: '' ifname: '10gige-agg-1 (lagg0)' devid: 1 phys_ifaces: 'all; no addresses' Coalesced by: --

Cause

There are not enough IP addresses configured in a pool's range to cover the number of interfaces which have been added to the pool.The pool may have the rebalance policy set to "Manual" in which case IPs are not moved back onto an interface automatically, and by nature of this setting must manually have a rebalance command issued.

Resolution

Verify that there is a pool without enough IP addresses in the range to cover the number of interfaces configured by running the below command and analyzing the output. In the example output below, we see one IP address in the range where 10 interfaces are configured in the pool: isi network interfaces list -v subnet1:pool1 - HDFS Name Node In Subnet: subnet1 Allocation: Dynamic Ranges: 1 10.10.76.80-10.10.76.80 Pool Membership: 10 1:10gige-agg-1 (up) 2:10gige-agg-1 (up) 3:10gige-agg-1 (up) 4:10gige-agg-1 (up) 5:10gige-agg-1 (up) 6:10gige-agg-1 (up) 7:10gige-agg-1 (up) 8:10gige-agg-1 (up) 9:10gige-agg-1 (up) 10:10gige-agg-1 (up) Aggregation Mode: Etherchannel (FEC) Access Zone: Zone-Prod-Cust (3) SmartConnect: Suspended Nodes : None Auto Unsuspend ... 0 Zone : zonename.test.local Time to Live : 0 Service Subnet : Prod-Customer Connection Policy: Round Robin Failover Policy : Round Robin Rebalance Policy : Automatic Failback Add a range of IP addresses to the pool to meet or exceed the number of configured interfaces in the pool, or remove interfaces to meet the allotment of IP addresses configured for the pool.

Support Cases