The Tektronix 2712 spectrum analyzer is a very good unit and it is definitely worth your time to fix and keep. The 2712 is part of a family of compact spectrum analyzers (2710/11, 2714/15) made by Tektronix. There are service manuals readily available and collecting and reading them will give you good understanding of the units. The biggest issue will be BOTH troubleshooting and the cost of replacement parts. While all of the other suggestions provided will get you going in the right direction, I suggest you look for a failure of the following items in this order:
1. Intra-board cabling may not be correct, use the service manual to confirm cable routing. I believe the first page of schematics in the 2712 service manual has a drawing of the modules and the cabling. It is not uncommon for techs to troubleshoot a unit and then when they find it is not economical to repair they just hurriedly put the unit back together and it goes to scrap. Sometimes cable aren't connected, may be damaged (bent pins), missing, or connected to the wrong port or board.
2. The A21 100 MHz oscillator board may not be putting out the 100 MHz frequency needed by the 2712's subsystems, reflected in the error messages you posted. With a scope or frequency counter, check J110 (level unknown), and J130 / J140 (about -4 dBm for either) for frequency and output level. If there is no signal, or it is weak or distorted (check by 100 MHz or higher input scope), then you will need to replace the assembly or component level troubleshoot the board.
3. The A18 Input Attenuator module has the front panel input and the internal calibrator signal fed through a TO-5 can relay used as a switch into the first attenuation stage. Try cycling through the front panel Level steps to see if an input signal is displayed at any point. There are five attenuators where one of them could be damaged from an excessive input signal, typically from an accidental connection to a transmitter. Worse case is you can bypass the attenuator, or connect the calibrator signal directly to the RF Mixer low pass filter input for testing.
4. The A16 RF Mixer has internal and difficult to replace mixer diodes that can be easily damaged by a high level signal. If the attenuator is set to zero dBm attention, a high level signal can pass through the attenuator and can cause damage to the RF Mixer stage. While you can repair the mixer, some components are no longer available (think mixer diodes) but you may be able to substitute a different part that may or may not provide the same level of performance. You will need some good test equipment to trouble shoot the mixer, but at minimum a 100 MHz scope may work, along with a signal generator capable of -30 to 10 dBm and a frequency range of 100 - 500/1000 MHz. If you have equal or better equipment and time, study the service manual and the operation explanations and then start with good troubleshooting measurements.
Best of luck, keep us updated as you go. Paul/WB6GHK