All seems fine except my online score took a dive lately because of ISP bullshit with bad routers and blocked ports, but at least my audit and suspension scores are still at 100%. Node is back up and running well now.
I am wondering however about the traffic and storage pattern. I am in europe however I see barely any traffic from the eu satellite.
I mean it’s continuous 150Mb/s on saltlake, us1 is about a 1/5th of saltlake, but eu is like… almost nothing. currently storing average 50GB/mo, where the other two have a few TB (two TB of which happened since yesterday)
Is that normal ? why no traffic from eu ?
Have you checked whether your node is fully vetted in EU satellite?
The traffic coming out of SaltLake is test data and it’s ignoring vetting status at the moment…
ok, good to know. Then it seems even more strange that I get zero traffic from the eu satellite. Specially considering how little traffic it sent my way.
Please make sure that the online score is greater than 60%, otherwise your node is suspended until the online score would be greater than 60%.
However, if the online score is greater than 60%, the usage depends on the customers’, not the hardware.
Do you have a free space accordingly the dashboard by the way?
And what’s your node’s version now? If the version is way behind the current one (1.105.5), then no ingress to your node until you would upgrade.
Thanks for the hints but like I said, I get constant 150Mb/s downloaded to my node by us and saltlake satellites (which apparently is unpaid test data… bummer), just wondering why eu is not doing anything when one would think I would get more utilization from eu, being in europe. I definitely have online score above 60, just saltlake dipped at 89 due to my current outage, others are 99% and 100%. I have 5TB free. The earnings.py script only shows that ap1 is 27% vetted, the others it shows nothing so it seems that means eu us and saltlake vetted 100%.
Surely you mean version 1.104.5 is latest, not 1.105.5
What is “almost nothing”? 30 day ingress for EU1 vs US1 is currently about 1:10, so if the amount of ingress you get from EU1 is about 1/10 compared to US1, then it should be okay.
The best thing you can do is check the logs to see if there are any errors there related to the EU1 satellite.
Also, keep in mind that currently (until 1.106 deploys) the bandwidth usage is only order bandwidth and not actually used bandwitdh.
You are both correct that I provided a wrong version, and accordingly https://version.storj.io/ the default is 1.104.5 and 1.105.4 is in the rollout.
We uses wording from the customers’ point of view, so “downloads” = downloads from the Storj network = egress from the storage nodes. “Uploads” = uploads to the Storj network = ingress to the storage nodes.
So you actually said, that you have ingress from the customers of US1 and SLC satellites, and little ingress from the customers of the EU1 satellite.
This is quite possible, especially if your node is far away from these customers. We implemented a dynamic load balancing (choice of 2), described there:
So, if the success rate is low for your node it’s selected rarely for these uploads.
You are correct. I get currently 150Mb/s ingress from us and saltlake. I worded incorrectly.
When I say I have almost nothing in eu compared to us and saltlake, I mean eu has used 100 times less bandwidth and storage than us/saltlake. Tens of GB compared to multi TB.
How could I be farther from EU customers than us ones ? I am in Europe…
Or you mean farther from eu customers than other nodes are ? Possible but that would be surprising, I am smack in the middle of Europe.
From the earnings.py output. Most data is marked unpaid.
The customers could be around the globe, not necessarily from Europe.
You can just check, how many nodes in Europe: https://storjnet.info, so it’s quite possible that your node is slower than many of these nodes…
This is due the missing reports from the US1 and SLC satellites:
So, this is just incorrect data, thus - incorrect assumption of the unpaid data (the script compares the reported by the satellites used space with the node’s used space and assumes that this is an uncollected garbage).
It still does not make much sense to me that I would be faster than us nodes but not eu nodes.
I mean ping times, latencies are always higher from me to US than europe. I ran iperf3 to several public servers all over the world and US stats are always much worse than europe.
That my node is just slower than most eu nodes does not make much sense to me as I have some (admittedly older but still) good server grade hardware (Gb ethernet, 32 cores, 384GB RAM ) and good connectivity (1Gb/s fiber). In the recent tests, I have continuous 150Mb/s from us and saltlake, and my server load is barely above idle (1-2%). Only thing I can think of is that somehow my connectivity to eu is bad.
Maybe I am just surrounded by very high speed nodes… weird.
Would be nice to find out the reason for lack of traffic though.
I do not know, may be there is less traffic than from the customers of other two satellites? Or is it possible that your node is still not vetted on the EU1 satellite?
this is actually doesn’t matter much. Your node should have a fast speed between the customer and the node. Of course, it’s limited by your cap for the ingress (downloads to your node).
Of course not. We want as much bandwidth from your node as possible. The limiting is not our goal, so there is no option to do so. Of course, you can use QoS or SQM if your router is smart enough… but generally - no.
I meant your cap from the ISP by the way.
I definitely do not have an cap from my ISP.
My node could easily saturate the 1GB line if storj sends it my way. Would be nice to figure where that 150Mb/s cap is coming from for me.
You likely have logs, so you can filter IPs which connecting your node for uploads from your logs (the log level should be info). You also may log what’s your OS see as a connecting IPs.
Then check the geoip DB in the internet.
However, I do not think it’s worth to check anything. All data and its usage are paid equally.
yes, I believe that it’s all accounted and paid.
I just wonder why there is no traffic from eu when it would seem most logical there would be. And that there could be a lot more.
As a side note, I think part of the issue is that “decentralized” storj is actually totally centralized. There is no market. There is only one customer to the multiple SNO: Storj. Therefore, Storj can dictate its terms. Which I accept, that is the system they designed and I choose to partake in it. I know that it is possible to run a satellite independently in theory, but in practice there are none except Storj, so there is still no real market because there are no multiple independent producers and consumers who each set their own prices.
I was hoping Storj would develop into a real fully decentralized storage market but it’s a different model. I think there is room for many different models to be successful at the same time, so I hope Storj keeps growing for sure.
Sia seems to go the “real” decentralized way, but there are other problems and still a lot of work to do for it to be mainstream and easy to use. We’ll see where it all goes, hopefully more projects pop up
Anyone can run an own Community satellite or join the existing ones:
And mentioned SIA just proved all of that, thus a little usage and adoption.
We want to have a full decentralization and do not lost a value to be a fastest enterprise grade end-to-end encrypted and globally distributed storage in the world and being an economically sustainable. Right now it’s not possible unfortunately, you still can chose only two.
But we hope that when the mentioned problems would be solved, we can achieve this goal and not sacrifice the full decentralization.
