velotraveler
commented
1 month ago Glad to hear you're using potraceroute, and that it works well in single-process mode.
The Traceroute class was not written with thread safety in mind, it's basically a port of the original C-language traceroute with some nice additional features. So your multiprocessing/multithreading results are not surprising.
I'm sure the script could be restructured to make it thread-safe, but it would take some careful planning. I'm not working on potraceroute at the moment - but if you want to submit patches I'd be happy to review and test them.
Thanks again for your interest in potraceroute!
On Thu, Jul 25, 2024 at 06:23:29PM -0700, hgonzalez007 wrote:
We are trying to generate the traceroute for more than 400 devices. The version with a single process works spectacularly, but when using multithreading, calling to class (Traceroute, probe(ttl) ) it appears that it does not work correctly. Some threads abort prematurely, which delivers incorrect traceroute information. The same behavior occurs when using multiprocessing(Pool)
— Reply to this email directly, [1]view it on GitHub, or [2]unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>
References
Visible links
We are trying to generate the traceroute for more than 400 devices. The version with a single process works spectacularly, but when using multithreading, calling to class (Traceroute, probe(ttl) ) it appears that it does not work correctly. Some threads abort prematurely, which delivers incorrect traceroute information. The same behavior occurs when using multiprocessing(Pool)