In <list-1729012@2rosenthals.com>, on 03/06/21
at 12:56 PM, "Paul Smedley" <ecs-isp@2rosenthals.com> said:
Hi there,
>I actually thought I'd done a similar fix for stunnel - as it's an
>obvious cause of 100% cpu usage - but you're right - there isn't a
>workaround present for it.
>Rather than implementing the hack, it would probably be best to use the
>latest libcx which has a wrapper for select().
FTR, on Massimo's system, the looping is in thread 3. This is the first
client thread, if I understand the code. Each service connection gets a
unique thread.
Thread 1 is the daemon thread, which does the select/accept and kicks off
the client threads.
Thread 2 is the cron thread, which basically does clean up.
We need to see Massimo's tunnel.conf to know which service thread 3
applies to.
volver a la lista