Hello - i recently updated all my nodes to latest version. I have a widget exporter on all my nodes, and i can see my total used space has jumped up about 10-20%. Im not sure if that data is wrong or real. And anyone else seen this? I know they changed the api:
Before: {"used":0,"available":0,"trash":0,"overused":0}
Now: {"used":0,"available":0,"overused":0,"allocated":0,"trash":0,"reclaimable":0}
But not sure why that increased “used”?
Thanks in advance.
in this version used space it is All, useful data, trash, reclaimable- all together.
Hmmm… That also means its unpaid data - making it harder for me to see WHAT data i am actually paid for. Back to the AGE old problem - why do we allow storj to use space they dont pay for.
Its not my problem they need space for trash, this SHOULD be payed.
I think this change makes it even more obvious - we are not payed for all storage that storj decides to use. Meaning its a payout at much less than 1.5usd per tb.
I dont see any problem here, you have in API trash amount and reclaimable amount, so simple calculation do the trick
Also you have Average used space reported by satellite, this is true space that you got paid for.
I guess this is the case of “if you want to be paid for trash… then choose to provide space to a project that pays for trash”. It’s just not Storj. I suppose if there’s a guy currently beating-a-dead-horse about IPv6 support again… we can beat this one too 
SNOs get $1.50/TB for space customers are paying for. If there’s some trash-handling as part of the platform I’m OK with it.
Yea i understand they have to have some kind of trash mechanic. But why does it fall back on our cost? how is it fair, that storj made a choice to keep data for longer, for their security, but we pay for that security?
I’m not sure what you mean: they offer to pay at a certain rate, and not for trash, and 30k+ nodes decided to accept that deal. It’s a decision: not some sort of burden.
If they offered $1.25/TB, and paid for trash, and you made exactly the same every month… would you feel different?
trash is only 7 days, after that it is lost for storj. Just put your node make more cleaning, then reclaimable will be less.
I don’t think that’s true anymore with hashstore, is it? With piecestore the trash was removed-from-disk after 7 days. But with hashstore it’s only prevented-from-being-reclaimed for 7 days… then it will sit around for a random amount of time until a compaction occurs. Which can be way longer than 7 days.
I’m still OK with having a trash system. But I was happier when that trash was for-sure-gone after a week 
I understand the point.
Let’s say a scenario:
I have 300TB of disk space. 200TB is used.
Of that 200TB, 15% is trash — that’s 30TB of trash — data Storj has deleted but is holding on my disk “for security purposes” during the grace/trash period.
That means:
At 30TB unpaid × 1.50/TB=1.50/TB=45/month out of my own pocket.I am silently eating. Money that comes straight out of my pocket in electricity, wear on drives, and lost opportunity.
The argument “just accept the deal” is valid in principle, but the deal was not clearly communicated in terms of trash consumption — especially with hashstore, where trash no longer has a guaranteed 7-day cleanup and can linger indefinitely until compaction occurs.
The goalposts have moved after operators committed hardware.
I hear ya. I just don’t see it as you “silently eating” anything, and there’s no “money that comes straight out of my pocket”. It’s something that you agreed to, happening as you agreed to it.
They haven’t changed payout rates for awhile now, and trash has always been excluded: the same docs with the rate cover garbage as well.
They definately move the goalposts every time they change payouts. But not-being-paid-for-trash has been the consistent message since launch.
compaction is every 12h. so dont see any problem here.
I have 370 tb used
only 11.5 tb from this is trash
and 16.2tb reclaimable. - this number is so high because i have broken file in some nodes and compaction is suck to deal with it. And @Alexey suggested workaround not work with write-hashtbl, because the same broken file it give error.
Compaction events don’t clean all trash. Every 12h a decision is made about what data logs are compacted based on tunable probabilities. So @HGPlays is right: trash can stick around way longer than a week - and in general a hashstore node will be holding onto trash for longer than a piecestore node.
Yes, but as far as i know, it clean it more HDD friendly, every day some part, not all at once, bunch of small files.
I never managed to fill a drive in 5 years, not even the 8TB ones. So, with or without trash, dosen’t make a difference to me.
You can look at it like this: the payout rate is variable. It can be 1.5$/TB or less, depending on the trash, garbage, logs, dbes, etc.
Which is why you can always format the drive and go on with your life. It’s not like you would go to jail for breaking your part of the contract. And you’d still probably be paid for some time of data “storage” before Storj’s systems would consider your node as disqualified.
Or I’d be happy to adopt his nodes! 
. I’d even cover their transfer: better than a graceful exit 
We will all get a share after their nodes are disqualified 
The problem is mostly default settings for hashstore compaction. After adjusting it a bit I am down to 3 percent reclaimable which is peanuts in my opinion.
But unfortunately, that requires the node to constantly move its content around to squize out all air.