I believe those limits are meaningful for the router, not for the storage rig, if you use the internet for something else and you see that the node interferes with your activity…
So you don’t need to set limits if you think your rig can’t handle the traffic. No point.
storage2.max-concurrent-requests:
I think another storjer had set this to a small number because they had a SMR hard drive that was having trouble
In my case, yep, they were mounted NFS mounts. I set it a long time ago but maybe I wanted to tell if certain nodes had problems with different levels. Or maybe it was during the surge in test data. or maybe even earlier when storagenode wanted to make sync writes (which are really bad over NFS).
Obviously whatever the reason was, I forgot.
1 Like
Yes, definitely SMR too, good point EasyRhino.
Re: Clean-up update
On another note, I guess my 48 hour TTL guess wasn’t necessarily correct. Would seem the delayed deletes/purging surges I observed every two days was the old free data being deleted.
This portends very well for b/w & ingress in the near future.
2 cents,
Julio
2 Likes
Should have named this also a bump to Degress Bandwidth - lol 
Had a nice part of the day deleting at about 1 Gbps! Woot!
Heh
2 cents,
Julio
amusingly, today, Apr 28, every one of my nodes is showing a similar spike in reported average disk space used. Even on a node that has had zero ingress for several days
I presume it’s either some wonky accounting by the satellite or a bunch of trash got undeleted.
That’s a false report. Wait until tomorow and the spike will be gone.
The data from the last day has always almost never been trustable. Just give it time, it will work out 
according to stats Storj Network Statistics - Grafana
storj rised about 0.5PB in last 5 days
5 Likes
That’s quite a lot
Impressive, if it’s accurate
Using some really fuzzy math… to add 500TB in 5 days… is a 10Gbps Internet connection pretty much pinned 24x7 for 5 days (if it was from one account)? Must be on Select… because I think Global SNOs would have noticed that activity.
1 Like
They have, see weekly graph here: ** Traffic Analysis for Th3Vans storj.dk Server030 **. I don’t have such a nice graph but my ingress is looking similar.
2 Likes
When we were getting uploads that size… during the testing period… we were seeing rates 5x what we’ve gotten in the last week (and Th3Van was seeing multi-Gbps rates - which we can still see in his Yearly graph). It doesn’t feel like we’ve been anything but slightly-above-normal lately.
But I don’t really know: as slightly-above-normal for 28000 nodes can still add up to a lot I guess 
My Nodes don’t really show much more ingress too. Even sometimes less ingress than before. I think most of it went to select
Edit: in the Month graph, you can see that something happens there
2 Likes
I Actuly noticed it, I got 5TB+ in this 5 days.
3 Likes
That’s interesting but it’s total used storage is compared to a month or two ago, it’s just recovered from some deleted data.
That’s very interesting. my nodes seem basically the same. And th3van’s charts don’t really show a change either. I wonder how you got all the data.
1 Like
Doesn’t Vadim have 100+ nodes? 5TB is only like 50GB/node then… or 10GB/node/day
I recovered after deletion this data. Last deletion was also big for me.
2 Likes
Look at his monthly graph. The last 2 and half weeks has much more traffic than weeks before.
Also, it looks like this traffic comes from EU satellite, so nodes in Europe can feel it better (win more races because of lower latency)
I think there is more than several new clients now…with >10Gbps links.
The significant bi-daily deletes may have just stopped today. If your node/nodes haven’t seen egress spiking for hours by a factor of 4 or 5 you might be doing something wrong.
I posted my ‘fuzzy’ math earlier. When they are all throttling the network, I have seen ingress 10x normal this past week (for several hours). Grafana can really give you some really great overview.
I purposely use ‘multiples’ of normal in respect of the last month’s normal bandwith, and when these spikes are not happening on the current network. So it’s probably not just small incremental growth, it’s a couple of big width bandwidth guys. It’s definitely a network workout.
What I don’t want to see with these new usage patterns are any satellites choking as they have in the past. So fingers crossed.
Then we are good to go.
4 cents,
Julio
1 Like
If satellites need to send their bloom filters on a regular basis, to keep trash under control, I wonder what we should expect to see steady-state? Maybe the network growing for about 48h, followed by a slow contraction over 12h, over and over? Like it’s breathing?
1 Like