Arduino NANO faulty EEPROM

Has anybody noticed that EEPROM can be faulty?

First I suspected bad coding. Then I suspected RAM overfill, so I deleted all RAM hungry parts, so that RAM usage was left at only 550bytes.
Still nothing. EEPROM gets Written at locations that code never uses, and gets read back false values. Randomly and seldom. I could not determine any pattern how errors occur. Reading of false values made me think that it could be a EEPROM fault. As during the next read from the same EEPROM location value is correct. Actually most of the times. And code is the same. It is not changed between the reads.

Don't forget that the EEPROM has a limited number of writes on each byte before it starts to fail. It's ~100,000 which you can easily hit if you're careless about it.

Code example? It may be an issue with cheap clones.

Has anybody noticed that EEPROM can be faulty?

No.

EEPROM gets Written at locations that code never uses, and gets read back false values. Randomly and seldom. I could not determine any pattern how errors occur. Reading of false values made me think that it could be a EEPROM fault

First I suspected bad coding.

So do I.

If you supply the code you are using to test the eeprom read/write performance others can help.

Solved.

Well. I managed to get different micro's from different patches and with different bootloaders. All behaved the same. EEPROM was randomly filled with garbage, along with correct data in correct position.
So it was certainly not EEPROM physical fault. And it was not too many EEPROM writes problem.

Big attention on code again. And there it was. There was a piece of code, that calculates addresses where to write small blocks of data. 18 bytes. Randomness was created because of unsigned int that was only declared and not zeroed before usage.