I'm seeing some strange errors in nginx.log about hosts trying to access an inexistent domain:
2014/08/29 14:23:18 [notice] 7155#0: *8578 no memcache server available, cannot post request, client: 126.96.36.199, server: webmail.deleteddomain.it, request: "POST /service/soap/ HTTP/1.1", host: "zimbra.host.name"
2014/08/29 14:29:11 [info] 7158#0: *8672 SSL_write() failed (SSL:) (32: Broken pipe) while sending to client, client: 188.8.131.52, server: webmail.deleteddomain.it, request: "POST /service/soap/WaitSetRequest HT
2.2.2:8080/service/soap/WaitSetRequest", host: "zimbra.host.name"
webmail.deleteddomain.it is the inexeistent domain. @deleteddomain.it was initially configured in Zimbra, but lately removed.
I checked cos, domain, users, server configuration and there is no reference to this hostname.
It cannot be a misconfigured client because there is no DNS resolution for webmail.deleteddomain.it so the client would neither reach zimbra server. I don't know where else to look for this, any hint?