Most data is 7 days TTL. So storj pays one week and uses our storage another two weeks for free. 
3 Likes
TTL isn’t trashed, it’s (supposed to be) deleted as soon as it expires (=the expire walker runs).
1 Like
Please check the satellite. Only Saltlake is a testing satellite, all other are customer-facing.
1 Like
If it is intentional why show a graph in the user interface that isn’t working?
3 Likes
@javierxam, Please advise how did you made such a nice graphs for your node(s)?
1 Like
It seems to be a customer who uploads his backups after a certain time. Bring it on.
9 Likes
I don’t understand the question. I can repeat it once again. The advantage of enabling the dedicated disk feature is that the node stops stracking used space. It simply skips that code path because its faster to not worry about used space and track just the free space instead. You simply need to understand what this feature is designed for. What do you expect to see on the graph if the node itself doesn’t even spend a single second to sum up the space on disk?
All the UI can show at this point is 0 byte and that is exactly what it is doing. It works as expected. If you want to see the graph all you need to do is disable the dedicated disk feature.
Thank you for linking the issue. I am going to close that with a comment in a minute.
1 Like
Nothing. I expect to see no graph.
No, it “works as currently implemented” but it’s not “as expected”.
The expectation is that if the data is shown it’s correct. If it cannot be computed, the graph needs to be hidden. It’s a bug to show graph with known bad data.
Edit: This graph has a lot of other issues, reported many times, I don’t know what’s up with storj and that urge to mislead with data visualization…
1 Like
I understand all that BUT how hard can it be to do a df and show the used space correctly …
1 Like
If that bothers you so much I suggest submitting pull requests to change it for your needs.
As an alternative you can also use grafana. I personally prefer grafana over the multinode dashboard. Anyway even grafana can’t display numbers that arn’t cacluated by the node in the first place. If I want to see used space on my grafana dashboard I have to disable the dedicated disk feature. Especially after finishing the migration to hashstore it is a lot cheaper to run the used space calculation now.
1 Like
How do you get the size of the trash folder with df? How do you know how much space on disk is used outside the storagenode folder? I can already smell 2 issues and both come to the conlusion that it would be better to show 0 bytes…
Edit: Oh and the fun continues. If you select a satellite on the dashboard should it still show you the result of df for each satellite? Now lets sum up the size of each satellite? Lets better keep 0 bytes instead.
1 Like
It does not; I just disagree that “there is no problem, works as expected, not to be fixed” approach of explaining away user reported bugs under the rug, when the correct response should have been “Thank you very much for pointing this issue out, we’ve triaged it, and assigned priority -9000.”
I.E. I fully understand that fixing display of bad/missing/improperly presented data in the dashboard nobody looks at anyway is at the very bottom of priority list, as it should be, but it’s a bug nevertheless, and the issue tracking it shall remain open until it’s fixed, if ever.
I (and I’m sure many others here) cannot contribute to your and many other codebases for various reasons, including contractual obligations, but it does not invalidate the existence of the problem.
No. Lets hide the plot. 0 is definitely wrong.
Is not this high school level CompSci concepts: Null (absence of data) and Zero (data value of zero bytes) are not the same things.
3 Likes
Complaining about storj dashboard is a waste of time. I guess the person in charge left years ago… 
5 Likes
The used space value is actually available via Prometheus endpoint, at least on a node that has dedicated disk enabled. It reports the exact same value as the OS would report as used space on the drive the node is using for storage.
I don’t understand why it would be as designed not to display this value in the dashboard if it is known by the node.
You won’t see the trash amount, but you can roughly determine that by comparing this used_space value with the used space value reported by the satellites - so node would be able to display both.
And I don’t code in Go, so won’t be able to do a pull request.
# TYPE used_space gauge
used_space{scope="storj_io_storj_storagenode_monitor",field="count"} 995102
used_space{scope="storj_io_storj_storagenode_monitor",field="sum"} 7.466598454680433e+18
used_space{scope="storj_io_storj_storagenode_monitor",field="min"} 6.84808091648e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="max"} 7.719498264576e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="rmin"} 6.848600080384e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="ravg"} 6.871560224768e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="r10"} 6.84978077696e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="r50"} 6.856491139072e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="r90"} 6.920979611648e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="r99"} 7.010036513505e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="rmax"} 7.010164670464e+12
used_space{scope="storj_io_storj_storagenode_monitor",field="recent"} 6.84808091648e+12
Again that doesn’t work. Try to select a satellite from the drop down in the dashboard. What is the node going to display? A simple df output isn’t enough to fill the dashboard with live. You would complain either way.
Disable dedicated disk is the way to fix this.
Wouldn’t it work to just use the os reported free and used space? So people can see some fancy graphing and won’t complain? The only thing they’ll miss is the used space by garbage.
and used space by satellite
1 Like
Yeah I was thinking the same, but the data is very inconsistent. If the Sat is not reporting anything, you have suddenly 100% trash. This will create confusion. So I would just drop that value then
Won’t complain? Is that even possible? Disable dedicated disk will fix it.
We are talking about a feature that you can enable. I can explain what exactly that feature is doing. It skips used space calculation in a few places. If you need used space values simply disable that feature. It is as simple as that.
Haha yes, I was a but too optimistic there.
But I have to say, with the new hashstore feature the calculation of used space is blazingly fast. If this would be the default in the future, there would be no need to do workarounds anymore. (At least as far as I can tell. The disk space calculation went from hours to just seconds)