HDD Space Allocation

Node Operators FAQ
,
1

Question for you all who know way more about it than I do.

I have a 10TB (labeled) HDD on my primary node. Storagenode says that my allocated limit has been reached. In my docker run command I have the following -e STORAGE=“8.0TB”. I selected this limit based on the fact that when I run “df -h” the HDD shows 9.1T of usable space. Then since it’s recommended to leave ~10% overhead, I rounded down to 8.0TB for the limit.

My question though is that my current df -h and df -H reports that my drive is about 80% full:
image

Then my CLI dashboard shows the following:
image

And my web dashboard shows the following:
image

As I believe I’ve read that Storj calculated the allowable storage space in TiB, so since my limit is set to 8.0TB in run command, I guess that lines up with the 8.0T used in the “df -H”. Although, I’m then confused as to why a 10TB drive would read out as 11T in “df -H”.

Then, I’m confused as to why the CLI dashboard says that 8.1TB is used, and then the web dashboard says that there’s >350GB disk space remaining (of the 8TB limit from the run command)…

Anyone have any insight to the situation I’m seeing?

Also, I guess I’ll go ahead and ask, with the information I’ve presented, what could I safely increase my run command limit to? I’m assuming at least 8.2TB and still have that ~10% overhead.

2

df -h

^ shows size in binary

df -H

^ shows size in decimal

3

Your timing is a little unfortunate as there is currently a lot of garbage collection going on that is seriously messing with the reporting. Everything moved to trash seems to be counted double after that move, which is why the CLI dashboard is displaying much higher use and going into the negative free space. Users have also reported that the node software is not following limits and instead of enforcing them more strongly like the reporting suggests, it’s actually letting more uploads in.

You should be able to safely set your node to 9TB, after all that HDD is 10TB. HDD manufacturers aren’t lying, they’re just one of the few who use the unit correctly. The OS will simply use the wrong units and display the TiB number with a TB unit. Storj supports both, but displays the decimal units. If you feel better about it you can also assign 8.2TiB, but it would effectively be the same thing.

Right now I advise you to keep it as is though, simply because of the weird reporting issues going on. Give it a week or so, then see what it looks like. You might want to give your node a restart at that point, because I think that might correct the reporting issues.

2 Likes
4

Appreciate the informative response. I’ll just let it run and see what it looks like in a week or two.