On Jul 15, 2:26 pm, "[email protected]" <[email protected]> wrote: > Hi there sorry if i wrote to a wrong place but could not find bug > reporting tool on the website.
Bugs go here: http://code.google.com/p/memcached/issues/list > I had unit tests in PHP that were sending flush_all and right after > that setting array of values. Right after that i was getting them back > in single get queries. Even that timeout on my items was 60 seconds i > was getting empty results. I have added sleep 1s between flush_all > command and set and everything went back to normal. > > Im not expert in C so i just had a quick look in the code and i think > it counts time in seconds, then invalidation of rows also happend on > second. But as i said im not expert in C and i dont know memcache code > so well. > > Could anyone please confirm is that a bug or my misunderstanding or > what could be other reason of this? This is a fairly accurate understanding, though I don't think I've seen anything that's done that lately. What version are you running? > Is flushing in separate thread? could it be delayed till after i set > my item so its not accessible? Flushing is completely lazy. It is decided that an object does not exist when the item is retrieved and is considered too old for various reasons.
