The ability to Gracefully Delete Nodes

I duno how relevant this really is, but from what i understand it could be a fairly useful feature for new / confused SNO’s and for the network to help keep data repair low.

In the example below, the new SNO created 10 nodes and figured out afterwards he only needs 3, this puts him in a bind, because he has no correct way to delete the nodes… or atleast not i’m aware of…

Conclusion / Suggestion
I suggest that GE modified so that Nodes can be deleted from the network correctly, without causing loss of network data, instead of getting 50% of their total earned held back amount they get… maybe 10% of their held back amount or whatever people think is reasonable.

PS. there should be added a warning to GE when running it before the 6-15month 50% payout mark, so that people will be aware that they forfeit part of their heldback amount for deleting the node early.

The reason i write 50% is because that GE (50%) + 6-15month (50%) = 100% held amount
in this case only the 50% paid for GE is relevant

rants and reasons below venture at your own peril
i duno if there should be some sort of financial benefit to this… but i think it’s unreasonable to think that people will wait 6 months or 15 months to remove a node they made by mistake.

in some cases there would ofc be no financial benefit to be had from a node doing a Graceful Delete, if the node is to new… but there would still be data on it … i mean a node of 3 months may not seen any real benefit aside from the SNO keeping a perfect record on not having nodes die…

But keeping a tally of node deaths for a SNO’s account might be useful, i duno…

basically my thinking is that there needs to be a way to remove / delete a node when one wants, and to create an incentive for a SNO to do that there should be some sort of reward… even if it’s not a financial one… ofc the problem with it just affecting the account is that if the SNO is permanently leaving the network, there is again no incentive to do a Graceful Delete.

Maybe just change GE so that if it’s done before time it’s not 50% held amount payed out, but maybe 10% ofc then one would have to add like some sort of warning.

I know this might not be a big issue, but i also doubt it would take to long to adjust the GE procedure so that it has two options that are basically the same thing, just with different rewards

okay, i think i’m done ranting now, and thinking this through… going to put the conclusion on top…

I love the way you have evolved in posting your replies ever since our last interaction :heart: They are short and to the point rather than transporting the user in multiverse (just finished watching Star Trek Discovery :slight_smile: )

I am impressed how you posted the gist of your post at the top and your reasoning at the bottom.

3 Likes

i really hope you had no votes left :smiley: and thanks i try… did attempt to cut them down to short form for a while, but that just ended up often destroying the over all sense of what i write, so if it’s still relevant when i finished writing a post, then if its a long post i add a conclusion which goes on top for TL;DR purposes.

but for posts like this, it makes very much sense to keep the reasoning, since the final result of the post may not end up looking like what i suggested, so if i want the eventual programmer to avoid the logical fallacies that i took into account, then the logical deductions needs to stay.

Not crazy about STD xD but i’m sure i’ve seen worse.

1 Like

Hi @SGC,

I’m not sure to understand your proposition.
You propose to allow SNO do a Graceful Exit before Month 6 (When the monthly held amount is 50%)?
I don’t get the thing about the 10%. Do you suggest to keep 10% of the SNO held amount if they do a Graceful Exit before Month 6?

no i suggest that 10% of the held amount is payed out when a node does a GE before the 6month / 15 month mark.

the GE and long term run time of a storagenode has a few factors in relation to the Held amount.

after 15 months of node operation a SNO gets 50% of the held amount paid out, that leaves 50% held amount which is usually payed upon successfully running a GE after reaching the current 6month mark or the future 15month mark.

thus that means a regular successful GE is worth 50% of held amount and thus i use this as my baseline, and then i just figured 1/5 seems reasonable…
not enough to make it worthwhile to shut down a node… but still might give some minor incentive compared to just destroying the node.

i want new SNO’s to be able to remove nodes they make by mistake because they don’t understand how the network works…

so 10% is paid out for doing a GE before the designated held amount payout time, that would make 90% of the held amount forfeit, thus leaving plenty of incentive for not deleting nodes before time… but in some cases there isn’t a choice, and just destroying the nodes is bad for the network.

data loss causes repair and repair is expensive… the only ones to pay is Storj Labs and SNO’s
i guess thats why we get half value for repair… it’s a 50/50 split lol of repair costs… because it’s basically SNO’s fault that the repair happens, but it’s Storj’s network so :smiley: 50 / 50 lol can’t believe i didn’t spot that sooner.

1 Like

If the node start and successfully finish the Graceful Exit after 6 (15) months it will receive 100% of remained held amount back.
Could you please mark those thresholds more precise in your proposal?
They can be understood in different ways at this point, as this discussion shows.

i adjusted it, but got to thinking that i might want to go read the GE documentation again to be sure… but the 6-15month mark isn’t really relevant for this…

i may have misunderstood something because i was of the belief that GE during the 6 month mark wouldn’t give 100% payout, because the 15 point mark for release of the additional 50% wasn’t reached… but i maybe wrong on that… will go read up on it when i got more time than now…

for now i wrote that i am referring to the 50% of the because thats the relevant part of the total earned held amount.
not sure how else i could simplify it… i’m open to suggestions, or corrections.
but if we don’t have a specific 100% mark, then anything is 100%… so :smiley: well % are hard… maybe i should use fractions instead…

the whole you add 100% and then add 50% and subtract 66% to be back at 100%
kinda has a confusing element on most people… maybe my mistake was using % at all

Definitely. The storagenode will receive a 100% of remained held amount after the successful GE.
The GE can be invoked if node older than 15 months. The age is temporary reduced to 6 months.

The partial return of held amount (50%) is happened after 15 months in the network.
So, if your node is older than 15 months, you will receive a 50% of held amount from that satellite. If you call a GE from it, you will receive 100% of remained held amount, i.e. remained 50% of initial held amount.

1 Like

The SNO could also be malicious and be trying to force out competition on his /24. Some thought should be considered about if we need to be nice, or be careful about incentivizing that behavior, because it is bad for the network.

Long rant short
yeah i fully agree that incentivizing malicious intent would be a bad practice, i just cannot see why this would do that.

rants and reasons
sure which is why i went for 10% of held… which means first you have to vet the nodes and then the first month is what like 75% held of which 90% would be lost if the node is deleted to soon, and 2nd month would be the same i think… can’t remember exactly … but something like that… so instead of 25% earned it would be +7.5% so 32.5% vs 25% now… and then is it the 3 or 4th month or something it goes to 50% so 55% if one leave correctly through a GE rather than 50%

so really i don’t think there is any real room for using it maliciously …
less than 10% and it almost becomes pointless, already pretty low reward compared to just waiting for the 6 month mark presently…

ofc with the 15 month mark it would become a bit more useful…but still 90% of the held amount would still be wasted… which is what after the 9 month mark i think its 3 months worth of payout’s that is held…

i would counter your argument with that repair is expensive… at the moment if one created extra nodes by mistake, there is no correct way to shut them down…

doesn’t even need to be a financial reward, could simply be that it would avoid some sort of black mark on the account… ofc what subtracts from that idea is that why wouldn’t one just create a new account, especially if trying to force out competition.

also it’s a question of how relevant thing it really is… i mean the need to destroy / remove extra nodes, most would figure out they don’t need them early on and thus they won’t have much data…

but still there should be an option to atleast make a node leave the network in a correct way rather than having to simply disconnect and essentially “damage” the erasure code or whatever it’s called, decreasing the integrity of the data closer to the repair point.

it’s wasteful…
but i don’t see why would would use it without a benefit… ofc some will, but if there was atleast some reward from it… then i suspect many will use it.

ofc my new nodes now almost earned 1$’ coming up on a month… and 10% of nothing is still nothing… i duno if this is needed… only storj labs can really answer that.
but in any case the feature should be there… even if there isn’t a “financial reward” from it.
if one can call 10% of held amount paid out from decommissioning a “newish” node as a financial reward, i suspect the actual amounts will be borderline nonexistent.

The point of the 6 month or 15 month minimum was to encourage SNOs to be in it for the long haul. That’s kind of undercut if you’re giving them back any percentage of held amount. And asking them to do a graceful exit without getting anything in return isn’t reasonable either.

There may be a small cost to having to do repair on the small amount of data, but you have to offset that to the lower reliability of nodes if exiting has fewer or less severe consequences.

All in all I’m not sure it’s worth it. You could offer the option for voluntary unpaid graceful exit for those who want to be altruistic. But then that has to be weighed against possible support tickets of people who didn’t read carefully and missed that they would lose their held amount.

3 Likes