Replicate error: Unable to connect to non-participating unit

Hi BrianL,

Yes, the Unit1-to-Unit2 replicate job has been working since May and automatically completed another incremental backup this morning.  I also had a Unit1-to-Unit3 job working for three years until it started failing Sep3 and never recovered.  I figured the incremental backups had overwhelmed the filesystem of Unit3 (Duo v.2).  So I reset it to factory default, re-installed Replicate, re-registered it (with you guys' help) and created the Unit1-to-Unit3 job that now complains 'unable to connect to Unit2'.

Regarding your suggestions below, we have no firewall issues: the LAN of units 1 and 3 has always been on plain-vanilla DHCP / NAT behind a cable modem.  No reason to set up port forwarding.  To 'remove Replicate user binded', I guess means to delete it from the 'Network' panel, right?  But if I remove Unit2 in that way, wouldn't that permanently disable use of the (currently working) Unit1-to-Unit2 backup?  Would we risk recovery in case something happens to production Unit1 while we are solving this problem (which should only involve 1 and 3)?

One more thing: back in 2012, Capone helped us sort through problems about serial numbers and license keys that originated when we switched unit 1 and 2 and their names.  Could today's problem still be some kind of leftover from that fix three years ago?

Thanks again,

replicant