Well on our way…
Here is my stats for the tests on 11 nodes, UTC+1
Maxed out my 1gbps connection

CPU doesn’t seem to care much about it.
Will update testing images when I notice it happening for today.
Same Dream Machine SE with 2.5Gb FTTH
IDS/IPS and TD enabled
I would disable ips/ids on storj traffic.
Nope. The security of my network comes first. Storj is an extra that takes advantage of hardware already online.
The last tests still have been with bestofn in different variations.
Now restarting with filterbest 50%.
Then you do security wrong, just segment storj from your internal network. Storj should not be able to talk to your internal stuff. And you decrease the load on your hardware. Win win.
Who is they here?
Slower nodes should not be expected to get as much data as fast nodes.
Forgive me, but I didn’t explain all my infrastructure.
Of course I have several VLANs!
- Default
- IoT
- Camera (no Internet access)
- Guest
- VoIP
- Servers
I have exposed some services on the wan, and obviously needed to be.
Everything that is private, or that I can, is under VPN.
P.S. IDS is Conntrack, and in my case it’s that layer that ‘you never know’.
Ah, Storj is burning my exos’s zpool with 2 raidz1 vdev
A few days ago it had settled to 150mbps
layer7 on encrypted storj data? Which wont work. And looking at your vlans i guess storj would be on your 6. Servers vlan? So it share vlan with rest of your servers? If you care about security you would put storj on separate vlan and turn off ips/ids for that traffic. Cause it wont do anything meaningful anyway.
Yes! I am now stubborn in bringing UDM to the limit.
In the future I will exclude storj, it’s just a tap from a mobile application.
Sure up to you! But if you care about your security put storj on separate vlan.
Not sure what you did differently, but two of those were the fastest I’ve seen yet. Closing in on saturating my 1gbps.
Looking forward to the explanation when you get a moment.
I certainly will! It is now running on the TrueNAS Scale. I am currently working on more important things, in the next round he too will be isolated.
Perfect anto294 good luck!
This time it did not go up as high
I now also have a graph of CPU usage by the node process (this is on a 4 core VM, so maximum would be 400).
EDIT:
What’s with the 10 byte size piece?
2024-06-04T13:26:17Z INFO piecestore uploaded {"Process": "storagenode", "Piece ID": "4YGJ7TEP5746DNLYUMOM65VZLSVR5I27QDG4IPGZFBU6A4I6SGSA", "Satellite ID": "12EayRS2V1kEsWESU9QMRseFhdxYxKicsiFmxrsLZHeLUtdps3S", "Action": "PUT", "Remote Address": "34.85.151.116:12400", "Size": 10}
Should’t small files be stored by the satellite directly? Usually the smallest piece size my node gets is 512B.
That is a different satellite. To be honest I don’t care for now.
I didn’t want to open a discussion on Microtik. Working great! I have an average of 4gbit/s. I wanted just say that storj testing is not helping in an already busy environment. Classic cherry on the cake.
I have already expressed my thoughts about this changes… We all need to work on bandwidth if we want to stay alive and fill that hdds!