4 ms·
There is. I'm surprised they didn't try it. http://msdn.microsoft.com/en-us/library/cc713687.aspx http://msdn.microsoft.com/en-us/library/cc713687.aspx
by tedunangst 15y ago
There is. I'm surprised they didn't try it.
http://msdn.microsoft.com/en-us/library/cc713687.aspx http://msdn.microsoft.com/en-us/library/cc713687.aspx
- jerf 15y agoThank you. I'm not a Windows dev so I didn't expect I would necessarily be able to find it very quickly. Reading that page it is clear that this is either the exact scenario that page is covering, or at the very least one of the core use cases. It's triggering my Wrong Thing To Do senses, but, well, when the GC isn't quite right, hacking around it may be the only option.
- sams99 15y agoReally? We register for notifications from the GC using a magic threshold number that could mean anything. Then we quickly notify the rest of the webs a GC is pending on our "message bus". They let us know they are safe from GC at the moment. If they are not you are in a pickle. Then we notify HAProxy that we are about to run a GC and ask it to tell us when it is done draining the current requests and taking our web offline. Once notified we perform the GC manually Then we notify HAProxy that we are done with the GC so it can add us back to the pool. What could possibly go wrong?
- JasonPunyon 15y ago2 magic threshold numbers that could mean anything :)
- tedunangst 15y agoOr you could just tell haproxy you're going to be low priority for a little bit and not worry about every single last request never being processed on GCing server.
- sams99 15y agoyou do realise that only about 0.09% of our requests were affected by this, catching them all is the whole point here.