Wow I rarely get more than 200mbps. Better location/internet/hardware or all of these things I imagine.
Also: aannd it’s gone for me.
Wow I rarely get more than 200mbps. Better location/internet/hardware or all of these things I imagine.
Also: aannd it’s gone for me.
also gone here, storj need time to setup it propertly
This is the normal test run. A few stages with increasing concurrencies and each stage takes about 100 seconds.
that looks like a big blip on my graph from earlier today. I see what people got excited about (briefly)
Strangely… I’ve seen a few egress blips over the last few hours. I don’t remember seeing testing grab files back off of nodes before?
We are going to start the day with a longer stress test today. Don’t panic. It still is just testdata. As soon as the test is running I can write down some details about the current test results.
Hang on to your potatos, kids!
![]()
I like roasted potatoes…
Yeah, I believe you are frying mine…
https://forum.storj.io/t/1gb-ram-is-not-enough-anymore/27531?u=snorkel
Meh. Seems like fairly average test throughput for me. Nothing special… ![]()
Not gonna lie, the tests are kind of fun to watch on activity graphs.
I honestly stare at mine too. I’m too excited to trim Chia plots in favor of making more space for Storj.
Found some more tests to run for today. At the moment we are fine tuning rate limits. So you will see various throughput tests over the day.
I’ve noticed considerably less “waves” of data since getting test data from US1.
The ingress data is hitting my servers at almost exactly 120 Megabits/second.
Testing is done. We will switch back to SLC now.
Now to the test results. We can see that the maximum throughput without surge nodes has increased by about 20% in less than a month. We don’t know for sure how much load these customers are going to upload. It is more like a range. The good news is we can cover this range with no surge nodes for the most part. Only on the higher end of that range we would have to mix in surge nodes. And the second good news is even if we go nuts on the possible range and take the most unlikely scenario into account we can make that possible with a few more surge nodes.
This has been an interesting journey. We started with an almost impossible target and a few releases later with a lot of code changes it turns into a simple math problem that is easy to solve for any customer that needs it. Now we just need to wait to find out if this plan works out. Sorry I still don’t have an ETA or any more details for you. All I can say for now is that we are ready for these customers. And thank you all for your patience and the additional capacity you have added.
Do you plan to continue the space reservation process until the deal is signed?
I thought that… independent of specific potential customers… that they were going to maintain 3-months estimated capacity at all times going forward. So even if their sales pipeline contracted… they’d still keep 3-months worth filled (just a lower amount). I hope so.
If you look at a 3-month customer-stored-data graph and squint a bit… is it fair to say paid data is increasing around 1.5PB/month?
I think the best answer for now is that SNO’s should expect similar data flows that they have seen these past few weeks. There may be starts and stops like before. I dont think there is an exact upper target for SLC data right now. However, you should see overall usage across all Sats to be comparable to recent activity.
It’s almost 2 months old now, but there were targets:
It doesn’t look like SLC has grown enough since that post for us to have filled all that space: so capacity-reservation data is probably still growing?