The local oscillator in "All American 5ive" vacuum tube
AM radios all drift an annoying amount at the upper
end of the AM BC MW band. The oscillator would be
running at about 2MHz, and warm up drift (from cold
start to about an hour being on) is typically 20KHz.
Enough to make that station at 1520 tune itself out.
AM radios used a hartley style oscillator using the
equivalent of a triode with its plate to B+, grid
capacitivitly coupled to the LC osc tank, and cathode
connected to a secondary winding on the LC osc tank.
Usually an air variable cap, and fixed inductor wound
on a cardboard coil form.

VFO's for ham radio work would involve higher frequencies,
and I would think that they not drift anywhere as bad as
the AM radios did. I looked at a few tube VFO schematics, and
I don't see anything that different from the AM radio
hartley osc circuit. So how did they avoid drift, or were
you expected to leave your VFO on all the time?

Slightly beter components, time to use and change some capacitors to
compensante for the drift, and some stayed on all the time. Voltage
regulation was usually used also.