1 Like
My valid partial percentage on chiahub dropped to 82% from the usual 99+
However, i do see that my pool share and estimate plot size are fine.
Most of the failed partials are reported as double submitted
It looks that for now the network is fine (no stalled partials, all peers on the same height). Also, it looks that all blocks from yesterday were processed today (Flex pool). So, except those stalled partials, everything worked fine, just got delayed a bit.
Looking at what happened, and what was suggested, there were potentially two different motivations behind that attack. The first is just for kicks, and that could be partially mitigated with payments on transactions. The second is (as suggested) that for a whale with good setup, it may be actually worthwhile to launch such attack to trigger those stalled partials on others, as reward during that time may be bigger than those micro-payments. Kind of scary.
1 Like
Yeah, well prepared solo farming whales. But I would not rule out OG pools either, whom I believe donāt rely on partials.
1 Like
Agreed.
We may not know, whether OGs were affected by that or not.
If you think that the main reason behind stalled partials was the nodeās overwhelmed up-speed, that would affect not just those partials, but all other non db updating traffic, as such also OGs. So, potentially only those with high upload bandwidth would benefit from such attack. The fact that we know about those stalled events, is because we saw those on our pool websites.
1 Like
True. Good point!
(20 chars)
Iād say its better.
However i still only have 40 peers not 80, and my ram usage is still climbing.
Not sure what to make of all that though.
I would restart chia.
If you didnāt change your config.yaml, you should see 80 peers (I think you didnāt). If that is the case, one reason that you donāt go that high is that maybe your node is somehow rejecting newcomers?
Maybe somehow your ISP is doing some extra filtering for you (as Chia wants to go broad)?
When I checked my peers yesterday, there was only one from US, and the rest was from all over the world (except Antarctic). To me, that is rather strange, as all those connections incur substantial network delays, especially hurtful for P2P connections (chatty connections).
Also, while 80 is recommended, that is actually not the whole story. When you reduce it, you will get less peers, but you will dedicate more bandwidth to those remining ones, so the end result is about the same (from network point of view). This is basically the main reason that a lot of nodes (like mine) got overwhelmed, as even with only few peers, my node was doing everything it could to choke the upstream bandwidth.
No, I didnāt alter peers, so would like 80 again.
But my ram peaked at 12.4 gb, my old box I just sorted with a new ssd only has 16gb, that feels a little to close for comfort to me, and Iām not buying more ddr3ā¦
Might just look at a cheap mobo and processor ddr4 gen for my new empty case.
Lol, fastest backtrack ever, found 32gb on ebay for 75 quid.
There was cheaper but I need 4 sticks, donāt think my board will take 2 x 16gb.