Bad NOS tubes...

In fairness to the vendor, any tube that tests well on a given day might be prone to early failure that cannot be predicted based on those favorable test results.  However, I do also empathize with you in that to have 4 out of 6 fail so early is way beyond the laws of chance.  So, I have to wonder about the state of your preamplifier.  Before you pop in yet more new and expensive tubes, perhaps you should have a competent technician check out the unit.  What you've experienced simply should not happen, even if the seller were a charlatan, and I am not ready to conclude that he is.  And by the way, a bad switch or volume control can indeed cause the symptom you describe, so did you check out those elements? So, check the pots and switches, and check the voltages at critical points in your preamplifier.

The driver stage is typically low stress, and, since there is no gain in the driver stage, most likely to be "quiet".  Try the "bad" tubes in the driver stage and let us know what happens.  What is the vintage and brand of the original set of "stock" tubes? 

Guys, Here are the facts as I see them:
(1) The OP bought 12AU7s for a preamplifier that presumably uses 12AU7s, so there is no question that he bought the right tube type. In terms of operating parameters, all 12AU7s are alike. (The OP should report to us if he changed tube types, but he did not say that he did.)
(2) A tube that looks good on a tester can only be relied upon to have been good at that moment. Tubes can fail unexpectedly and prematurely, whether NOS or current manufacture, but the hooker here is that 4 out of 6 tubes appear to have gone bad in such a short time. That defies the law of averages for tube lifespan, when tubes are used within the recommended limiting parameters.
(3) For this reason, we suggested that the OP might want to check out the voltages in his preamplifier; one cause of such a catastrophe would be abuse of the tubes, for sure. They DID last for more than a month before showing signs of early death.
(4) Against this hypothesis, the OP also tells us that his set of OEM tubes, presumably also 12AU7s, are all working fine after many more hours of use in the very same preamplifier. Moreover, he re-installed the OEM tubes and they are sounding fine.
(5) So, if we want to exonerate the vendor, one more hypothesis remains tenable: Perhaps something went bad in the preamplifier subsequent to the installation of the new tubes, something that killed 4/6 of them prematurely. If this is so, then the OEM tubes ought to go bad too, after a while. To test this hypothesis, I advise the OP to continue listening with the OEM tube set for a month or two longer; see how that works. Some or all of them will fail as well, if the preamplifier has developed a power supply problem. (This could involve either the filament supply or the B+ supply.) OR, take the unit now to a good technician for a check-up.

Peace. Out.

Czarivey, For a source of truly NOS discrete transistors that may have been discontinued, try James Electronics in California.  They had every transistor I needed for rebuild of my Beveridge direct-drive amplifiers (ca 1980 vintage), when no one else could supply them.

Jea48, Gold-plated pins are a sure sign of recently manufactured tubes.  Vintage tubes of the good old days rarely had gold-plated pins unless specified by the type. For example, some milspec 6900s had gold-plated pins but most did not.  GE 5-star tubes sometimes had gold-plated pins. Other than those two instances, I have never encountered true vintage tubes with gold-plated pins.  And the pins are not usually made of steel.  They're probably brass or other copper alloy (which is why they oxidize and may need cleaning) or rolled "sheet metal" in most cases, although which tubes have what pins is an interesting question.  Just not steel.

Jea48, First, I apologize for sounding so pedantic.  Your post actually prompted me to do some research on the metals used in tube pins.  In the URL you provided, are you referring to this passage?: "The 6DJ8 family was originally developed by Amperex. The first tube in the family was the steel pinned 6DJ8 followed by the gold pinned 6922 and 7308. These latter two were premium versions rated as 10,000 hour life tubes. The 6922 was a premium industrial version and the 7308 was the top of the line guaranteed low noise version. Frankly, I’ve measured & listened to a lot of these and there is no correlation between model # and sound quality."

Based on the brief bit of researching I did do, my first thought is that this guy is incorrect as regards the composition of the pins and that he used the word "steel" without really thinking about it.  Moreover, I think any one of us would say that no tube ever made has "gold" pins (as he describes them, carelessly); at best sometimes pins are gold-plated.  Pure gold would be a terrible choice for a tube pin, because it is so soft. You cannot easily plate steel in the first place.  So the notion of gold-plated steel does not make sense, either.  And finally, this is a piece about the 6922 and its many related tube types.  These are frame grid tubes that were only developed at the very end of the "vacuum tube era", mostly for use in instruments and TVs.  So generalizing from the 6922 back to the tubes from the golden era of audio tube technology is dangerous, even if some 6922s do have "steel" pins.  But it's a good question, and there is not much solid info on the internet.  I do believe that older octal based tube types, like 6SN7s probably have brass pins that were plated, maybe with nickel. My NOS 12AX7s and 12AU7s seem to have some copper containing alloy that is tinned.  If I polish a pin with my Dremel tool, it eventually reveals a copper-y color.