xmrig - duplicate shares since yesterday

[Syracc]

Hi all,

I've got a strange problem with xmrig which seems to persists since yesterday. I'm mining with two computers, one has an AMD Ryzen 5 2600X and the other one an Intel Xeon E5-1650. Both are mining monero with the CPU. The Ryzen one also mines Etherum with an AMD RX 580 and gminer.

Some minutes ago I recognized that I've got like 26% rejected shares due to duplication. Researching a bit, I found out that theses could be happening because the mining software is outdated - so I updated the systems to the current version of xmrig ( https://github.com/xmrig/xmrig/releases/tag/v6.13.1 ). But still the error perseveres.

Because the error happens on my on both of my systems I excluded the possibility of an hardware problem on my side. Also because the error happened with two versions of xmrig ( 6.13.1 and the old one 6.12.2 ) I would likely excluded a software problem from my side too - maybe there was a system upgrade at prohashings pools?

Anybody else experiencing this? I'll try now version in between 6.13.1 - 6.12.2 maybe this could do the trick.

Greetings

Edit: I've forgot the obvious, both maschines are running Ubuntu. The Ryzen computer 20.04.2 LTS and the Intel one Ubuntu 18.04.5 LTS.

Edit 2: The error seems to happening less often with 6.13.1 as with 6.12.2 but still some duplicated shares are shown in the logs.

[ToeKnee]

I am experiencing a high # of rejected or problematic rejected shares with mining XMR as well; I am using the latest stable version of HiveOS. Everything was working fine until about Saturday, July 17, 2021, then I started seeing problems. Prohashing team, check out what's up with XMR on your end to at least eliminate that possibility.

[Steve Sokolowski]

This issue has been resolved. Thanks for your feedback!

The cause of the problem was that RandomX requires recalculation of different work for every person. In yesterday's release, we changed the calculation of work for other algorithms back to once per block using a different method, which reduced latency in calculating work but also introduced the bug for RandomX customers.

[Syracc]

Thanks for the quick fix! My % of rejected shares keep dropping - logs seems fine so far.

I've opened also a ticket for that, I'll close it now.

[Setvin]

Was wondering about this myself. Looks like we are clear to start again. I notice a lot of pools avoid rx/0 and im guessing it's because it requires a lot of upkeep or just a lot of people aren't knowledgeable about it. Really grateful you guys decided to add it.