What log do you mean?
I doubt that.
It should be the synthetic data mentioned in the other thread:
US target 5 PB
logging entries for uploads/downloads and other stuff
My logs showing uploads/downloads in realtime, so I donât understand what the problem is⌠
Few upload log entries per second, while ingress near 80-100mbit
I can see a good number of max. size pieces uploaded atm, so this might be the reason for having not so many log entries?
However, without doing the exact math we can not know wheter it fits or not.
Most of it is probably paid on repair traffic to redistribute the data the leaving nodes had.
2 Likes
Yes, there is a lot of repair traffic atm. Also healthy pieces numbers looking strange:
Hopefully those numbers are as wrong as any other statistics.
No. You got it backwards. The held-back is to incentivize nodes to stay. So is 6 month requirement for graceful exit.
Stable long term nodes are much better. Each node leaving generates pressure on the network. Each year node is staying contributes to network stability.
Therefore the nodes cannot be made profitable. The payout rates that make it worth buying hardware will attract wrong people and destabilize network.
The point of the project is to make use of underutilized capacity, remember that. Itâs crucial to the project success.
Otherwise itâs just another datacenter. Why storj needs you if they can buy hardware themselves and cut out the middleman?
5 Likes
According to storj stats only 6% of no longer active nodes did a GE, the remaining just got disqualified for whatever reason. I am sure the benefit from all those held backs overcompensates the repair costs.
There are also healthy nodes that donât bother with GE without being disqualified.
For example, I have a few nodes occupying free space. When i need to get some space back I just destroy the dataset of one or more nodes. Waiting a month to get a few bucks makes no sense to me to bother with GE
In that sense yes, I agree with you, the current heldback system is a poor incentive.
I feel if there was no heldback amount more people would do GE because of a stronger incentive this being a right thing to do. But once you offer compensation â this becomes financial transaction, not social, and the held back amount is not enough to bother.
Another aspect is that people are more averse to losing than not gaining.
I.e. not getting $15 feels less painful than losing $15.
Hence, I would prefer there to be a deposit made by the node operator to escrow, earning interest, that is lost if they donât do graceful exit.
This would not have worked in the begiiign when storj needed to attract node operators. But now there are plenty of them, it may be a good strategy going forward to attract people who are serious about hosting nodes.
3 Likes
may the fct that nodes sruggle with updates because the service is unresponsive while quiting for updates not work?
my node did this this morning.
Can they? Hardware is sold for real money not storj token, I would be surprised if they had that much real money.
Ooh more test data incoming. Just finish migrated 1 of 3 nodes to LvmCache - my body is ready!
2 Likes
wow you seemingly have one of every possible disk format, I like it.
What is âreal moneyâ?
Anything of value can be used as money. If storj token has value â it can be exchanged for potatoes, credit, currency, or servers. Storj token is as real as any other form of currency. Just like currency, it exists because people believe in its value and promise to pay.
4 Likes
Did somebody say potato?
1 Like
i think potatoes as in âraspberry pi storj nodesâ
1 Like
itâs not easy to implement from the legal point of view, so I do not believe that it could happenâŚ
I do not like the naming of a low power devices as potatoes, however, if it can power the GladOS, why it cannot power the Pi?
4 Likes