Doh! I’d run out of disk space! So if you get the error
SelfEncryption Error A generic I/O error.
It would be nice to have something that better matches what the situation is! 
4 Likes
Throwing out there that AVAST antivirus also flags safeup. Did allow it to run, but it’s been “sent to their thread assessment team”
Not sure what, if anything, can be done about antivirus flagging the software.
edit: and safenode-manager
2 Likes
here is a few more for you
"ubuntu-9.10-alternate-amd64.iso" b4a1be6412cc89957e11c9bd095324242baad8ce3788c981683fd143bd1477e4
"ubuntu-10.04.1-desktop-amd64.iso" a45e4c597059ce3441fd9156d9106244c25292f175043474cffe21bf3c320fe3
"ubuntu-23.04-desktop-amd64.iso" 80d88c666263718d9ee271acdf38710065d4cc235837eed7fa7314f0a53d8080
"linuxmint-21.2-mate-64bit.iso" d239d34f77f73f4dc60ef3e8a12ffc87ded440dd886536d5aaaea0b2c15e2963
"linuxmint-21.2-cinnamon-64bit.iso" 36dcf903ccd7d4ff67c097eb4d62063bed6be8448baead8d86c3fa0548bb8aa1
looks like a big 100% success rate for uploads iv not had a single problem apart from running out of SNT
7 Likes
about to go to bed but last observation of the evening is the network has a heartbeat
approx. every 5 min processor usage goes up from 15%-30% to 70%-90%
so if you are running nodes make sure and keep some CPU head room up your sleeve
8 Likes
It seems common, happened to me when made a Maid price tracker app, had to send it to Kaspersky for verification that it was ok.
2 Likes
Does anyone know what a bad node is?
I did a bit of poking in github couldn’t really figure it out and posts I have seen are vague at best.
Is there criteria, a benchmark if so what? Is it a comparison to close nodes? I have no idea.
Kinda feeds into my next question, is there any thing in bad node detection that helps prevent too many nodes for available resources?
I fear that when the masses come running crazy amounts of nodes under-resourced is going to cause trouble.
5 Likes
D:\a\safe_network\safe_network\sn_cli\src\files\iterative_uploader.rs:116
What is this mystery location on D: ? (when everything happens on C:)
Ran a few more analysis on 50 nodes within 1 container/LXC:
Observations:
-
35 out of 50 nodes failed to bootstrap with the 1st peer id in network-contacts (likely due to out going connection error + handshake timeout)
-
15 out of 50 nodes did receive tokens while still having handshake timed outs on outgoing connection errors
-
% blocked peer ids is mostly < 1% for those 15 out of 50 working nodes
-
Incoming connection error rate + handshake timed out was 4% where as for outgoing connection error rate was 95% based on unique ip/port combinations
-
If determining % outgoing connection errors with handshake timed out based on unique peer ids from all unique peer ids combinations → its 76% that experienced this issue
-
If determining % outgoing connection errors with handshake timed out based on unique ip/port from all unique ip/port combinations → its 95% that experienced this issue
-
In the case of outgoing connection errors + handshake timed out, only 2.34% of all unique IPs/Ports were associated with my own nodes, the rest were other folks’ nodes
Clearly these errors aren’t stopping my node from earning, but would the node(s) have earned far more if these weren’t happening to begin with.
It seems if others are blocking my node, and vice versa, time is getting consumed here to detect and block (its TBD on why certain ports are showing up that I never entered as input parameters into the safenode when slicing on just my own WAN IPs). Also, since the pid is continuing to run, and its ‘observed external address’ is also changing, it can keep being marked as bad, reset the state, and then marked as bad again, repeating this process for a really long time, as long as its observed external port changes by others (I assume)? Is that a fair statement?
If we weren’t trying to access or send data through ports that say were never opened say on my firewall, then perhaps these handshaketimedouts wont be happening? Is that a fair statement?
Note: UPnP is disabled on my router, and only 200 ports are open specific to these safe containers as far as I am aware.
Just curious why the blocking rate is so low in an environment where the outgoing connection errors are affecting 95% of all unique ip/ports communicated with… (not complaining here, but is it because we are focused on ‘errors’ or ‘criteria’ with incoming connection failure rates more than outgoing connection failure rates?). Or, is this because the teams’ current code base requires repeated error conditions to be triggered within a certain amount of time for a block to really take place… and if so the thresholds just aren’t being hit (not necessarily a bad thing here)?
I am also curious if MaidSafe’s droplets and other folks nodes, all have a unique WAN IP per safenode pid (because many are running it on the cloud as is), or a vast majority are sharing a WAN IP and running multiple safenode pids behind that single WAN IP?
The initial network-contacts list seems to contain all 50 unique WAN IPs.
Is MaidSafe also noticing handshake timeouts (against their own peerids for any of their 2000 safenode pids backed by say the 50 WAN IPs)?
Just thinking out loud, is this scenario I am experiencing more likely an issue when running multiple safenode pids behind a NAT with say a single WAN IP?
Also noting here, for many months, I ran at best 1 safenode per WAN IP (I have multiple WAN IPs) and at most maybe 2 to 5 safenode pids running simultaneously…, but I never tried 100s of safenode pids behind a single WAN IP.
Its only recently I am testing this stuff out at larger scale from home, and in the back of my mind, I keep thinking “is it my environment or is it the network’s code base”, or both at fault (an odd combination of potentially user error + current code base causing the incompatibility/issues noticed above). I will keep investigating on my end for sure… and I hope I get to the bottom off this issue… 
.
Thank you all for providing feedback on your setups and other vital details along this journey.
Update: Trying to grasp what you said earlier regarding this section of code base:
9 Likes
observation of the evening is the network has a heartbeat
approx. every 5 min processor usage goes up from 15%-30% to 70%-90%
so if you are running nodes make sure and keep some CPU head room up your sleeve
I wonder if it’s specifically node ‘heartbeat’ or a close group one. If that’s the case it may be a hangover of the nodes that Maidsafe started all being started at once. It might therefore disappear or get evened out once the network is bigger and has run for a bit.
If it’s genuinely a node ‘heartbeat’ then it might be smart to start nodes at an interval of minutes that isn’t divisible by 5.
4 Likes
I am fully aware of the irony! Professionally embarrassing… Like an F1 driver scraping a car while parking at the supermarket. He’d never hear the end of it.
My only defence is as a conscientious storage admin I try to minimise my allocations to VMs to save capacity.
4 Likes
ValentinesNet 24:41:07 hrs for 154,486 chunks
Made payment of 0.253248899
Made payment of 0.044616818 for royalties fees
BasicEconomyTweaks 04:26:42 hrs for 106,593 chunks
66245 already existed in network (from previous tests)
Made payment of 0.713781553
Made payment of 0.125912849 for royalties fees
Current test significantly more expensive than ValentinesNet, but much faster upload
6 Likes
I think so.
Currently:
Close nodes
- Repeated Connection issues
- Bad price quotes offered
- Failure to replicate data( thresholds?)
Non Close nodes
- Connection issues
- Reported as bad by close nodes
All still being worked on and up for improvement and debate!
We’re mostly looking to get data to base more decisions on I thnk. If something is looking sus (or we know is wrong), then we can work up criteria for it. I’m not sure why you’re seeing such handshake trouble.
all have a unique WAN IP per safenode
No, we have 40 nodes/droplet atm.
4 Likes
Sorry mark, missed this
DialError(DialPeerConditionFalse(NotDialing))
Hmmm, very odd one. Means it’s not attempting to dial because the conditions that dictate if it should are false…
I’ve no idea why though. The condition it notes is that we’re not already dialling. But your nodes dont appear to be. If you can gather more logs (SN_LOG=all), and try starting more that’d be great. And we can take a look next week 
4 Likes
I am interested, how do the other nodes know a close node gave bad quotes?
This is under the title Non Close nodes.
Again I am interested what is meant here as it seems contradictory at first
Also can I ask what the equation that is being used to determine the nanos to quote? Please, and I think others would love to know in the opening post of each test/betanets.
1 Like
Would be fun, if the wallet could just be imported in the new client after deletion or just a “clear” command to erase all clogged up issues.
Initial impl is basically: tell us everything you factored in. We rerrun the calculation check certain things (records stored etc) are increasing compared to prior quotes. (@qi_ma can offer more insight)
Note: That one is not in the testnet yet, but it is in main (waiting to be tested)
Non close (to me) nodes. We ask their close nodes what they think of it.
I’ll let @qi_ma expound on that as he was last about there!
5 Likes
scott@scott-OptiPlex-3050:~$ safe files upload -p /home/scott/Pictures/Screenshots/screenshot.png
Logging to directory: "/home/scott/.local/share/safe/client/logs/log_2024-03-29_08-32-48"
Built with git version: 01c2e57 / main / 01c2e57
Instantiating a SAFE client...
Trying to fetch the bootstrap peers from https://sn-testnet.s3.eu-west-2.amazonaws.com/network-contacts
Connecting to the network with 50 peers
🔗 Connected to the Network "/home/scott/Pictures/Screenshots/screenshot.png" will be made public and linkable
Splitting and uploading "/home/scott/Pictures/Screenshots/screenshot.png" into 3 chunks
⠤ [00:00:06] [----------------------------------------] 0/3 Retrying failed chunks 3 ...
Unverified file "screenshot.png", suggest to re-upload again.
**************************************
* Uploaded Files *
**************************************
Among 3 chunks, found 0 already existed in network, uploaded the leftover 3 chunks in 12 seconds
**************************************
* Payment Details *
**************************************
Made payment of 0.000000000 for 3 chunks
Made payment of 0.000000000 for royalties fees
New wallet balance: 1.000000000
Keep getting this error trying to upload what does it mean?
1 Like
I’m delighted to share that despite my non-technical background, I’ve successfully navigated joining my first testnet using a Raspberry Pi 5, complete with an over-sized SSD. I’ve established connections with 20 nodes without much difficulty!
14 Likes
@happybeing any idea why Vdash does not report my records? I have 783 of them in the folder. I’m on Ubuntu 22.04.3 LTS.
4 Likes