High CPU usage caused by the Hazelcast client about hazelcast-python-client HOT 3 CLOSED

@Svedrin
Shouldn't we increase the count to reduce cpu usage ? When count is one, the thread will wake up from the loop more frequently. Right now it wakes up every 100 millis. (1/10 of a second). For cpu, that should not be thathigh. When I run an idle client , it consumes at most 1% of a core in my machine. What is the cpu usage you are seeing ? If it is high, then can you describe the how you are using client ? May be we have another bug that cause this.

Integrating async-core loops does not really seem possible to me.
Is it done for similar libraries ? Do you know an example for that ? May be if we can find a smooth example that does this, we can offer the same thing.

from hazelcast-python-client.

@sancar

Right now it wakes up every 100 millis. (1/10 of a second).

actually it does not, because the timeout is set to 0.01, which is 1/100 of a second. asyncore calls this method ten times in a row, which means that asyncore itself does come back only after 1/10 of a second as you're saying. But it did ten syscalls in the meantime, whereas a single one would have the same effect here.

So setting count to 1 (or omitting it completely) and setting the timeout to 0.1 (which is what you're trying to achieve) would mean that you only use a single syscall instead of 10.

And well, the HZ client is already using asyncore, so it's integrated already :P

from hazelcast-python-client.

Pr looks good to me.

We'd appreciate being able to have the HZ client use our main loop ?
By integrating I meant this one. How can we expose an API so that, HZ client can be part of an external loop.

from hazelcast-python-client.