fuzzbench icon indicating copy to clipboard operation
fuzzbench copied to clipboard

Fuzzbench silently stops measuring runs when SQLite database is "full"

Open addisoncrump opened this issue 8 months ago • 0 comments

When the SQLite database is on a disk which is approaching being filled, it will cause Fuzzbench's dispatcher to emit the following error:

Process Process-1:
Traceback (most recent call last):
  File "/usr/local/lib/python3.10/site-packages/sqlalchemy/engine/base.py", line 1880, in _execute_context
    self.dialect.do_executemany(
  File "/usr/local/lib/python3.10/site-packages/sqlalchemy/engine/default.py", line 733, in do_executemany
    cursor.executemany(statement, parameters)
sqlite3.OperationalError: database or disk is full

However, after this point, the runs continue. In some cases, it seems that SQLite's detection of the fullness of the filesystem is overly cautious (e.g., this occurred at 91% usage... of a 3TB drive), and also that Docker seems to interfere with the fullness detection in some way.

regardless, this is a truly hard error that should cause Fuzzbench to terminate all runs early. As of now, it silently stops measurement but the runs keep going, which made this error very difficult to find.

addisoncrump avatar Jan 30 '25 18:01 addisoncrump