The GSLB pool is providing an incorrect IP to end users

If there’s no upstream DNS or NAT device, then the behavior is almost certainly coming from the GSLB logic itself. In many setups, disabling a pool member does not immediately remove it from DNS responses if the service still considers the pool available. Cached responses are a big factor here, especially with long TTLs on wide IPs.

Another thing to check is whether the pool has fallback enabled. If all members are down or disabled, the system may return the GSLB or wide IP by design. Also verify monitor status, not just the enabled flag. If the monitor still shows green, the pool can be treated as healthy.

I’d suggest reviewing the wide IP configuration, fallback settings, and forcing the member offline. Then test with fresh DNS queries to avoid cached answers.