A water bath does not really do it, it just makes the resulting accident site wetter.
There are containment systems that can control and redirect the escaping gas in an explosion and contain any shrapnel, but they are spendy and slow down the tank fill process. Due to space considerations they also tend to reduce the number of tanks that can be filled at once.
One approach that works pretty well and is reasonably cost effective, but not always practical in some dive shops is to have the fill whips on one side of a block wall or piece of suitable steel plate and the fill valves and pressure gauges on the other. Then if a tank ruptures, no one will be in the direct path of the blast or fragments.
It is exceedingly rare (literally a several million to one occurence) for a properly hydro tested and visually inspected tank to fail at any pressure below the test pressure, so it will be interesting to hear what caused this tank to fail. My suspicion is that:
1. Someone screwed up in the hydro testing or visual inspection process, or
2. The tank was subjected to excessive heat - as in a fire or in a refinishing process involving heat over about 300 degrees F.
Number 2 is a subset of 1 as a refinsihed aluminum tank should always be considered suspect and the the owner questioned in detail about the refinishing process used. If any doubt exists, a new hydro test should be done.