… and 43TB of data used is by far a personal record!
Fortunately AT&T didn’t complain about offending the unlimited data that I pay for.
(although I was using an Oracle VPS and I DID get hit with a bandwidth charge there for exceeding 10TB and that was kind of expensive)
43TB last month? Or was it July/August?
I moved my nodes to AirVPN, and recently moved them back to oracle.
I highly suggest configuring alarm on oracle so you are notified before it goes over limit. I’ve setup daily egress alarm for 300GB – if I have a few days exceeding that – time to move back to AirVPN.
The 43tb was a period straddling the month, mostly august some july.
and good idea on oracle alarm, I reconfigured things so it shouldn’t really exceed the 10TB allowance even under test duress… but an alarm would probably be prudent.
Oh hi there,
nodes meant to be set and forget, huh? ;>
(it allusion to our conversations about SNO payouts propositions)
Me choosing to mess with my nodes does not in any way mean that it’s required. I like to mess with things.
Yes.
My server in Wolfsburg is forgotten.
But my nodes at home are behind cgnat. So I have to use Oracle too
oh You “Chooose”, i thought You have to watch some limits or something, or things will go belly up, but ok then!
No, you try to bend the narrative but it does not work.
I could have set them up with AirVPN and have unlimited traffic and not worry about it. Setup is trivial and “just works”.
I chose to use Oracle, with limited traffic, because I wanted to understand iptables. It took me almost entire evening to figure that out, and I consider that time well spent. Bonus - Oracle datacenter is physically closer to those nodes than wherever nearest AirVPN endpoint is running at. I feel it’s better for storj. AirVPN has some weird peering and latency does not correlate with their claimed location.
I don’t have to watch anything, I will receive an email in the future if traffic picks up to exceed 10TB monthly, and switch wireguard interface on two nodes back to AirVPN. Hardly an inconvenience.
So, what are you trying to say? That it’s hard to run nodes? That they are not set it and forget it?
Edit. The better argument you could make — “you said nodes meant to be set and forget but you spend exorbitant amount of time on this forum arguing over trivial things”. To which would respond that — I neither run nodes, nor hang out on this forum for money. It’s entertainment for me.
Ok. I offered explanation of why they are. I have also offered anecdotes of how my other nodes work. You still claim the opposite.
At his point you need to either point to the flaw in my argument or provide supporting evidence to your point.
I’m convinced that if you have to spend any time “maintaining” your nodes — you are doing it wrong.
If you believe otherwise — then tell me what is it that you have to do? Because I can’t think of anything. Storj is responsible for node software. Operator is only responsible for providing power and port forwarding. Do you mess with your ports? I’m seriously curious.
i have to be ready for unpredicted situations, to intervene.
By my fault, or beyond my control.
sooo i have to be available to counter problems that may occur in case of offline time.
not that it needs me daily to do something, its jut not self-play all the time, i smile.
okay there is one thing i do every day regarding my nodes - i give some time for thoughts, “is everything okay with that bastards today” and then, “le me check”
Me too, I started this to tinker on stuff, not to get rich.
The end of brutal test traffic has made me sad because there’s no admin work to be done any more. No installing additional hard drives (always fun), no debugging and repairing broken nodes (mixed levels of fun but always interesting).
I wouldn’t mind getting rich off it, of course.
I wonder how you managed to get charged for something that is “always free”? I never registered a paid account and it just stopped working when the 10TB limit was hit.
Traffic is not “always free”.
I have converted my account to pay-as-you-go. This is to avoid the free instance being suspending for inactivity (the vpn server and a few podman containers I run there are very lightweight), and any other hurdles that can be resolved by throwing money at the problem. In other words, I don’t want anything to break just to save couple of bucks.
This is much worse. I don’t want it to just “stop working”. I want to notify me, and then I decide what to do.
This is very vague – what specific unpredicted situations are you talking about? Give specific examples.
And besides, hardware related mishaps are not specific to running storj – you already had a server running anyway, you would have been doing all that “handling of unexpected” without strorj.
Sure. Which is negligible in the grand scheme of things. I wrote the whole replacement node updater for freebsd, but it was a one-time deal, and I don’t touch it since.
Bingo!. Stop doing it. You don’t have to do it. If something goes wrong you will get an email. Don’t check “if maybe something is wrong”. Instead, get notified when something actually is.
Just now we lost power for 30 min. I got a slew of emails that two nodes and other services are unreachable. And then emails that they are again ok. I did absolutely nothing about it, because power company also sent me notification. I did not bother reacting other that reading it when I had a spare minute. If storj was not there – I would have done exactly the same things.
nah, i mean, an intervention because something F’ed up on my home server is not taking the same time as intervention when node needs me, those are separate events, different in time, so a node makes my home server potentially more my ass calling for help,
from unexpected things, over course of 5 years what had happen isssss that moment when some free DDNS provider failed causin havoc among SNOs, at first nobody knew why Online time is going down for so many … i mean c’mon, alex can F’ something up in code and then i HAVE to CLEAN the mess, like typing commands everywhere to forget decomisioned sattelites,… cmon, “unpredicted situations” are vague by definition … i can try to predict Alex will place some code to detect forum litigiousness and my nodes will get less traffic, so let me shut up then. :>
I understand your point, but let’s consider each issue separately:
I claim that node does not need any special attention, see below why.
This is not a problem of storj, or caused by storj, or only impacting storj. it’s a free ddns provider. That’s why I refuse to use anything free, including oracle “free” tier, without becoming a customer. As a counteranecdote – in my whole decade of using cloudflare I have never once had my dns fail. Maybe because DNS is one of their core businesses… Would I use noIP/duckDNS/other crap? Never. Won’t even consider. Storj or not, I want to be able to access my home server.
No you don’t, and that’s the deal. You let storj run node software, and it’s on them to maintain it. If they leave a lot of crap around – not my problem. They will just get less space. So they are incentivized to fix issues.
I have never had to do that. And yet somehow magically every node has only 4 folders under blobs. Magic. Or maybe storj doing their job. Maybe if you waited, instead of rushing to fix things, and gave storj more time there would have been nothing to fix.
Right, but I never experienced any, and wanted examples.
Lol. I would not care if my nodes get less traffic, or more traffic. Again, this is not my problem. I was asked to run this piece of software, and make it accessible on some port, and provide wallet where money drips every month. I have done all that.
Storj is implicitly incentivize to make node as efficient as possible (and this coincidentally benefits me too) so I don’t have to do anything.
My only monitoring is that TCP port is accessible. That’s it. I promised this to be accessible and to keep it accessible does not require me to do anything, just keep the server and modem powered. Which they were already.
Lets consider more examples of unexpected:
so… as long as my server exists it is going to be running, and there is nothing for me to do.
The key is to not self-inflict issues, like using weird DDNS provider, or crappy router, or crooked hardware, or flaky cables.
OMG no, im not even reading this. 
Sorry, Cheers.
I also converted mine to pay-as-go a few years ago, because I couldn’t even get one of their free ARM servers in their availability zones until I did. I didn’t pay any money for those years up until I exceeded the 10TB limit.
I want to tell my anecdote too. I running nodes since 2019 and changed something like once a year. I do monitor them via Multinode, but only to confirm that I can see the same behavior as other SNOs here. If I wouldn’t be here on the forum, I even wouldn’t bother to install a multinode dashboard.
Excuse me, but why should I do this?
Or this
It’s none of my business. I’d like to help you solve your problems, not block you.
I do not have even that. I want to be a normal SNO, who managed to install nodes following our documentation. If I see some lacks, I will submit an issue.
Of course the external monitoring would be better, but it’s more for the huge setups, not my tiny three nodes with 9.28TB of the shared space in total.