Yes, mine to. It’s only on this specific machine. But I can’t figure out what makes it different….
So, I could ask a query every 5 sec to keep alive?
any message, you could even send a ping though im not sure we expose it.
yeah i dont think its exposed through client api
https://github.com/EventStore/EventStore/blob/release-v3.9.0/src/EventStore.TestClient/Commands/PingFloodProcessor.cs#L74
Added a query every 5 sec, made things even worse, disconnected every minute.
Tried running server 3.8.0, same behaviour.
Not even close to understanding this behaviour. Somehow the socket is closed, regardless of heartbeats and traffic.
I’ve minimised reconnect time so they can use the system without too much interference. But I’m not sure what the tell the customer. ‘Your server sucks’ seems a bit vague…
/Peter
Hi,
I had very similar problem. EventStore was just silently dropping connection with heartbeat timeout. My app has lots of read/writes operations (ReadStreamForward, ReadStreamBackward etc) and at some point it was just hanging up. (e.g. AggregateRepository was trying to restore aggregate by calling ReadStreamForward and this method was was just blocking entire app). I resolved this problem by replacing Wait() calls with non blocking Task.Run(…). It’s working now but clearly i couldn’t find the real reason of this behaviour.
субота, 16 липня 2016 р. 09:46:23 UTC+3 користувач Peter Hageus написав:
Sounds like task pool exhaustion.
I considered that for a while (the service is single core vm), but it happens regardless of load.
/Peter