If it came apart with JB, i think that the prep. may have been flawed.
A friend has a Baja bug type Volkswagen , with a highly modified motor. For the previous build of that motor, that went about 170 thousand miles, he had radically ported the heads, and in the process, he ran through the casting of one of the intake ports, leaving a small hole, which he repaired with JB weld. Now I know that an intake port would be cooled by the gas air mixture and all of that, but I am still impressed by the fact that that repair never failed. For JB or any of the epoxies, I think that one of the big issues is how clean the surfaces are of any contaminant, and that that is the probable cause of some of the glue-in failures.
For those of you that think that gluing and screwing is the answer, I leave you with this. Years back, before a match at Visalia, I asked a group of top shooters the glue-in vs. pillar bed question, and after every one had expressed his opinion, Lee Six, said that doing both was what he preferred, that he thought that it was better than either one by itself. Later, a Hall of Fame member, who builds rifles, and was part of the original discussion, told me that that may have been because the epoxies in use at the time (for glue-ins) was not as reliable as some of the newer adhesives like JB.