On 21 Feb 2020, at 20:21, Michael Gerstacker <michael.gerstacker@googlemail.com> wrote:
> On 18 Feb 2020, at 06:10, Michael Gerstacker <michael.gerstacker@googlemail.com> wrote:
>
> Once the consensus diffs are processed the load drops to normal.
> After some time without anything noticeable for me in the debug logs the CPU suddenly jumps to 100% again and stays there till another consensus diffs are arriving.
>
> Its not as worse as it was the first two days where i had 100% CPU load half of the time but i still have this about 10-15 times a day for a few seconds or minutes each.
> For the next few minutes after the CPU dropped to normal the throughput is close to zero so this cant be good for clients.
>
> It would be nice to have a fix in 0.4.4 stable or earlier so that i can decide if i want to buy a Windows license key or rather shut it down.
Does setting "DirCache 0" in your torrc resolve the issue?
Yes it look like thats working.
I lost the Guard flag (like expected) but in the last 24 hours i had no problems anymore.
Ok, thanks, I've opened a ticket to implement this workaround,
if we can't solve the underlying issue in the 0.4.4 release.
There's a suggested patch on the ticket:
https://trac.torproject.org/projects/tor/ticket/24857#comment:39
But we've had trouble getting people to help with Windows development
and testing.
Can you compile tor from source for Windows?
i've never done that before. If there is a step-by-step guide it might be interesting to learn.
Alex has put together a guide for cross-compiling for Windows on Linux:
Or if we get a fix merged into tor, can you install the Windows
"Tor Expert Bundle" ?
This should be no problem.
I think i will keep the Windows setup at least for some time so if i can help that way feel free to send me or point me to anything Windows related what needs some testing in real world.
Great, I've made a note on the ticket.
T