Steps to reproduce:

1. Run discovery, ensure that host has been created;
2. Edit hostgroup prototype so that it is longer than 64 chars: rename the "Test/vSphere/VMs/{#SITE}/{#SEGMENT}/{#CLUSTER.NAME}/{#HV.NAME}" to "TestTestTest/vSphere/VMs/{#SITE}/{#SEGMENT}/{#CLUSTER.NAME}/{#HV.NAME}";
3. Run discovery again, check host groups;
4. Check the server log.

Result:

1. Host's groups are not updated, an error in server log:
   {{ 230:20201122:111241.128 [Z3005] query failed: [0] PGRES_FATAL_ERROR:ERROR: value too long for type character varying(64)}}
   {{ [update hstgrp set name='TestTestTest/vSphere/VMs/DC0/CLS01/DC0_CLS01_Test/tst-tst01.tst.testtest.test' where groupid=55;}}
   update group_discovery set name='TestTestTest/vSphere/VMs/{#SITE}/{#SEGMENT}/{#CLUSTER.NAME}/{#HV.NAME}' where groupid=55;
   ]
2. Sometimes I see another side effect with a long hostgroup prototype: a host is created that has all host groups except the longest one. Can't reproduce this all the times.