The MX records for nauta.cu currently resolve to mail2.etecsa.net and mail3.etecsa.net, but the IP address that Mandrill tried to resolve in your example was 181.225.231.50. That IP was tied to mx2.etecsa.net, but is no longer in use by nuata.cu. So when Mandrill first tried to deliver your email to nauta.cu, it was likely during the time when the old MX record (181.225.231.50) was still cached, and before the new ones had fully propagated.
This is probably resolved now since the TTL on nauta.cu's MX records is so low, but Mandrill's servers are likely retrying delivery for some time until the emails ultimately bounce, and then they'll stop trying to send those same emails. For example, when a receiving SMTP server (nauta.cu's in this case) returns a 4xx type of error, most sending servers (like Mandrill's) will retry delivery for a set period of time before giving up. You'll probably end up having to manually re-send these emails if you want them to get through, instead of waiting for Mandrill to bounce them.