There is no parity interrupt on the 8085 or 8086/8088.

If you mean a memory parity interrupt, that is a function of system design, not a function of the particular microprocessor involved. Generally, a memory parity error is fatal, so one would typically place it on a non-maskable interrupt, such as TRAP on the 8085, or INT 2 (NMI) on the 8086/8088. This assumes, of course, that the memory parity error does not just crash the processor.

Parity error indicates bad memory. Parity checks compare the memory read with what was writen.

No. When adding new memory, you need to match what is already in your system. Parity modules have an extra chip that detects if data was correctly read or written by the memory module, depending on the type of error. However, a parity module will not correct the erro

Non-parity memory is memory without parity. Parity memory is memory with extra bits, sometimes one, sometimes more, that accompany the word. These extra parity bits are generated to a known value, typically to make the total number of bits on that word even or odd. When the word is retrieved, the parity bits are compared against what they should be. If they are different, then one or more of the bits in the original word or in the parity bits must have changed. This is an error condition that can be trapped. In a multiple parity bit system, the calculation of the bits allows not only for the detection of a changed bit, but also for the identification of which bit changed. This is known as ECC parity, or Error-Correcting-Code. Often, you can detect and correct any one bit error, and you can detect, but not correct, any two bit error. Since random bits changes are rare, those that do occur are usually one bit errors, making ECC parity valuable for high reliability systems such as servers.

A parity error always causes the system to halt.