Re: [tor-relays] Worse throughput with 0.4.8.x, on a slow CPU

Subject: Re: [tor-relays] Worse throughput with 0.4.8.x, on a slow CPU

From: George Hartley via tor-relays <tor-relays@xxxxxxxxxxxxxxxxxxxx>

Date: Tue, 09 Jan 2024 21:45:28 +0000

Cc: George Hartley <hartley_george@xxxxxxxxx>

Delivery-date: Wed, 10 Jan 2024 02:20:44 -0500

Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=lists.torproject.org; s=2022-eugeni; t=1704871237; bh=sbdZr4EF/d52FMOYf7kuKOI7VYVpfSE+tSBtoxc7WRA=; h=Date:To:In-Reply-To:References:Subject:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To:Cc: From; b=fjxtzZITO59In+ovmAXcVXgcyBzThNK9QgULRMO9Gm84QjKOpx3YLdC9AMyRx00fk 6f8NR5r2ZTbOX0MqUTcZtjX1q/1r/5tko45PeUrXC5pqy8IPg8nLnBAZ7YbeM37sPb vCZk5GTmV6FFpDGgvTxpqKSjjtLKNHlaInpONhEJnEv7/Dr6FFIAFn/e3jJ9wcOdzg bdU+x+thH7divIzLxaRTsrkGkkFp/NCxvZ96Y6E1U+TzNFxbGqsAv3ruJ+xpQaYZfs VRJMwtFzlN+sqYonREdKQIv7H0h63l8JYe1UnqiqHbqpvCOtj+GIWcQDXBjRVLW4b2 VJh5mF2ws0cSg==

Feedback-id: 68993610:user:proton

In-reply-to: <facb87e4-ed24-4cb0-bfca-10270edec676@mulb.us>

List-archive: <http://lists.torproject.org/pipermail/tor-relays/>

List-help: <mailto:tor-relays-request@lists.torproject.org?subject=help>

List-id: "support and questions about running Tor relays \(exit, non-exit, bridge\)" <tor-relays.lists.torproject.org>

List-post: <mailto:tor-relays@lists.torproject.org>

List-subscribe: <https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays>, <mailto:tor-relays-request@lists.torproject.org?subject=subscribe>

List-unsubscribe: <https://lists.torproject.org/cgi-bin/mailman/options/tor-relays>, <mailto:tor-relays-request@lists.torproject.org?subject=unsubscribe>

References: <20231213191546.068323bf@nvm> <8a0832a5-70eb-43d4-a4be-a42da3f0af04@torproject.org> <facb87e4-ed24-4cb0-bfca-10270edec676@mulb.us>

Reply-to: tor-relays@xxxxxxxxxxxxxxxxxxxx

Sender: "tor-relays" <tor-relays-bounces@xxxxxxxxxxxxxxxxxxxx>

Dear Jeff Blum,

Yes, I am seeing something similar on 0.4.8.9 (and potentially earlier versions as well, not 100% sure when it started). I upgraded to 0.4.8.10 today hoping it would go away, but I'm seeing it again. Watching in nyx (screenshot of bandwidth graph attached), reliably every ~30 seconds.

This is just Tor doing zlib-compression on some documents, you can somewhat combat it by assigning more cores to your machine.

Tor is mostly singlethreaded, but OnionSkin decryption, zlib compression and some other operations utilize all cores detected.

Regards,

George

On Monday, January 8th, 2024 at 9:54 AM, Jeff Blum <jeff@xxxxxxx> wrote:

On 12/13/23 06:15, Roman Mamedov wrote:

On the older version it gets about 80+80 Mbit total in+out. On the new one the
average is at most 45+45 Mbit. There are frequent periods where the bandwidth
drops to 5-10 Mbit for 3-5 seconds, while all Tor processes continue to use
100% of both CPUs, then gradually climbs back up.

Does anyone notice anything similar?

Hi Roman,

Yes, I am seeing something similar on 0.4.8.9 (and potentially earlier versions as well, not 100% sure when it started). I upgraded to 0.4.8.10 today hoping it would go away, but I'm seeing it again. Watching in nyx (screenshot of bandwidth graph attached), reliably every ~30 seconds, I see the bandwidth briefly plummet and the tor process CPU spike. Guard relay running in docker under ubuntu server on a Ryzen 3600 machine with 32GB RAM. Note that when the relay restarted after the upgrade today, it didn't do this for a while (maybe an hour or so? wasn't watching it the whole time), but then started glitching every 30s. Once it starts it does this every ~30 seconds forever. Relay has been running like this for weeks, maybe months.

best,

-jeff

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________ tor-relays mailing list tor-relays@xxxxxxxxxxxxxxxxxxxx https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays