Yvan

I agree with @nikic, I got all my servers stuck the same way (after a flush of APCu and a Varnish ban), resulting in many concurrent requests. 99% of the...

Good question :-) I don't know, but if you have a file descriptor, you can see what it's doing. At least that was how I detected the issue was in...

@mhouriet the fact is that along with the crash of #259 that was happening during deletions (iterating on many keys and deleting a lot of keys), my second issue was...

What was the process `53650` doing? As you can see, it's the blocking one: ``` futex(0x7f7d4ee8b094, FUTEX_WAIT, 53650, NULL) ``` But the other big issue to me is that you...

Maybe the fragmentation did this. As I've detailed in https://github.com/krakjoe/apcu/issues/127#issuecomment-244362339, fragmentation is not an issue in general use cases, but if you're near the memory limit, I guess that APCu...

@rcanavan the PID may change if it's another lock issue. For example you may have 1234 that's locking every other thread, then when the lock is released you'll have 2345...

@nikic thanks, I was mistaken on this one. I have `1 week, 3 days, 11 hours and 8 minutes` for a PHP-FPM started `10d 11h 9m 34s` ago, so that's...

@nikic you're right, that's not a PID! It's the file descriptor of the PID you're strace'ing. So if you're running a strace on PID=1234, and the futex says 5678, you...

@nikic I've just installed 5.1.12 (without `disable-rw-locks` argument for compilation), and still have big issues with `APCiterator`. If I run cache invalidation requests concurrently, everything goes down in a few...

An old virtual machine was still using 5.1.9 with the `disable-rw-locks` option, and crashed the same way. So RW locks shouldn't be the culprit.