This implementation constantly “bumps” lock expiry time while computations are being made. This should fix the situation when lock was mistakingly acquired by concurrent tasks because computations took more than a lock timeout. This supposedly lead to import task applying some updates twice, and maybe for continuing with updates application process after a fail (race condition).
This implementation constantly “bumps” lock expiry time while computations are being made. This should fix the situation when lock was mistakingly acquired by concurrent tasks because computations took more than a lock timeout. This supposedly lead to import task applying some updates twice, and maybe for continuing with updates application process after a fail (race condition).