To begin it is necessary to indicate the conditions for the safety of data. In each segment, SNOs must save:
29 pieces - critical, if less, then the file is lost
35 pieces - optimal minimum quantity for guaranteed recovery
80 pieces - this is the amount the client loads (maybe 89, but I don’t remember exactly this value).
By the way storj said that there are no lost files. If the network is so reliable, then I think the company will agree to pay me 1 million $ or tokens for the lost file? If you are not afraid of course…
You can immediately see that after 1 month even a new satellite has to restore data for safety. HOWEVER, server-side Europe does not. Maybe I’m out of luck and this is purely an accident?
Test point2 - may for saltlake and june for eu-north - 2m shift from start.
And here it is clear that only Europe north did not care about data integrity.
Hmm, maybe this node is just surprisingly out of luck? Ok, we’re exploring a few more.
And so, nowhere is the recovery process going. Thus, it is obvious that data integrity is not ensured, which means it is exclusively a reservation at an extremely low price (less than $1/TBm).
I think you’re just a troll. Every post I read from you is about you not trusting STORJ or that you want to leave the network or you have no manners if you ask for help. Why are you still here?
Here only facts about strange EU-north patterns. Storj said that’s for customers, but how only one satellite can work without repair, when other active use this tool?
I have. Because I have been holding v3 nodes since April 2019 and I see how other satellites work perfectly.
Rather, these are the last sparks of my stay with this project. I believed in his success, but burned every time. I do not disclose some things, but the community should know some of the data. Let everyone draw their own conclusions.
Sorry, but I see this member is not a native English speaker, and that’s all. He tries to explain his own emotions in English but not excellent.
But your aggressive post showing you as a hater to other forum members, I personally not like it.
If you think that:
a) file loss is normal
b) The situation when the support service can’t take one elementary action and they continue to deduct money from me because of THEIR error.
Then yes … I have no words, only emotions.
P.S. this does not apply to the topic. After all, we want to see only facts, not emotions, right?
I checked on my node (only been on the network for about 2 months) and my results are similar, no repair ingress from europe-north.
However most of my ingress comes from that satellite and I don’t know why someone would download data if it’s just a reservation.
Your reasoning makes sense and I think the developers might (and should )enlighten us on why there is no repair ingress for that specific satellite.
Let’s assume you’re right and bad nodes were not weeded out between saltlake and eu-north launch. What if it’s a large scale test of a different approach? Storjlings said it multiple times - repair cost hurts the network a lot, so maybe they want to see how much longer they can go without repairs if they use different parameters for erasure coding?
Ok… please stop using the word fact for what is clearly rampant speculation. Then stop jumping to conclusions that are by default always negative. Provide your evidence, then ask what could be causing that. That’s how we can have a productive discussion and avoid all the negative responses.
So let me try to translate. You’ve compared satellites and noticed that europe-north-1 has far less repair traffic than others. Though I’m not sure why you’re comparing different months on a node that started at the same time on all satellites. The repair listed on those payout overviews is repair egress, so that would be directly related to how old the data on your node is, not how old data on the satellite is. You should be comparing the same months. Regardless, your finding still holds up. So now you’re left with a question. Why is there less repair on europe-north-1? And that should be the end of the story, because it’s the end of your evidence. This means the topic title is already not even close to a fact (despite misuse of that word).
You can also absolutely not conclude files are being lost. Or that europe-north-1 is only used for space reservation. I’m pretty sure gitbackup is actually running on this satellite. So that would be a false statement just based on that.
I may have some more information that could be relevant though. Last I heard the repair threshold on satellites was set much higher than the initially documented 35. I believe it was 55, but don’t pin me down on that exact number. The last report also was that not a single segment on the network ever had less than 50 pieces remaining. The repair threshold was partially raised to thoroughly test the repair processes, but I’m sure also in part to better protect data. I would also like to reiterate that europe-north-1 is not a customer facing satellite. It’s used for testing and internal projects. It’s possible that europe-north-1 has been set up with a lower repair threshold. Maybe that one is actually set to 35 and so it takes much longer for repair to be triggered. 35 should still be a safe setting though and wouldn’t lead to data loss. I don’t know if a lower setting is what’s causing this difference. And neither do you. The only way you’re going to get an answer to that is by asking the question and not by jumping to likely wrong and always negative conclusions. So I kindly ask you again to try and be more constructive in your posts.
I would also like to dispute that this is merely a language barrier. While that clearly plays a role and I’m absolutely sympathetic to that, the choice to not ask questions but make up negative conclusions is not a matter of language, but a matter of attitude. And I would really like to avoid having to discuss that again so we can actually talk about the subject matter instead of attitude. I hope I managed to do a little bit of both here.
That’s a plausible explanation, I agree that it was a big jump going from “europe-north doesn’t have much repair traffic” to “europe-north is used as reservation” and that the post should have been formulated differently.
I guess we won’t be sure until we get an official response from the developers.
Hello, @BrightSilence! Thank you for the high-quality response, an good answer!
These numbers were obtained directly from the satellite via uplink. I received them during testing the network as a user. However, now I need to double-check this.
In any case, with equal storage volumes, the difference is anomalous and this is clearly not casual.
By the way, why we have 3 test satellites (north from your words)? Still alive stefan, saltlake and the north.
The numbers I mentioned were presented during the InfluxDB time series webinar, if you want to look it up. That’s a few months old now though, but I haven’t seen repair behavior change since then. So I’m guessing the settings on the customer facing satellites are still the same.
isn’t north a new satellite anyways… so maybe it doesn’t have enough damage to start repairing lots of stuff yet… i suppose it take a good while before repair really kicks into gear on a new satellite…
so yeah, don’t really see the point … ofc maybe somebody forgot to turn on repair lol stranger things have happened…
From having read lots of good constructive posts from @donald.m.motsinger, I can you assure he is not a “hater” by any stretch of the imagination. He was simply pointing out that OP has a real negative attitude.
I do however agree that language barrier often leads to misrepresenting one’s actual attitude, so yeah, we as a community should still always try to stay neutral/positive towards other forum members.
