I found other topics with similar discrepancies but didn’t see any resolutions. Every month the storage value reported under my Payout Information is always quite a bit less than what is actually being used. The disk utilization value and graph on the default web page is generally pretty close to reality (sometimes it has been off as well but given time it almost always fixes itself). From those other posts it sounds like the value comes from a bloom filter on the satellite servers but why is it never accurate for me? My uptime is generally just shy of 100% (99.80+). It feels like I’m getting shortchanged every month by not getting paid for the actual amount of storage that storj is using on my hardware.
$ du --apparent-size --si -s storage/blobs
That command confirms the reported average disk space used value and also that the Disk Average Month value is not accurate.
Average Disk Space Used This Month is reported by satellites, it consist of submitted and settled usage, confirmed by the customers and your node. On nodes it showed as an average, but reported daily.
Used of Total Disk Space is a used disk space calculated locally.
Usually they should be pretty close, if your node submit orders in time (orders expires in 24h).
There is also could be a problem with Garbage Collector, it sometimes cannot finish its job (for example, when databases become corrupted or inconsistent for some reason or if the disk subsystem is too slow) and your node may store the unpaid data due this problem.
The problem firstly was noticed on systems with corrupted databases or BTRFS filesystem.
There is a discrepancy between the reported and actual disk space usage on your Stroj node. This could be due to issues with the Garbage Collector or database corruption survivor io mod. Check your databases, optimize the disk subsystem, and keep your node updated.
What about Disk average month. Is it calculated locally or reported by satellite?
Because at the end of the month Average Disk Space Used This Month and Disk average month should be equal right?
Had a difference of about 5% last month.
Disk average month is a different representation.
Average Disk Space Used This Month is every day of the month till today but Disk average month is the average of the total months including the days from today to end of month (720 hours). These days have “currently” zero used so you will have to wait till end of month to be correct. It will increase daylie.
Both. The actual values received from the satellite (daily), then calculated locally to show on the dashboard for the selected period (sum).
So it uses the same information from the satellites as an Average Disk Space Used This Month (average).
These values may not match even on the end of the month, because they calculated differently. The value in the table is more accurate.
Thank you for the responses. The server is an 8 core 16 thread AMD Ryzen (should be plenty powerful enough) and the HDD is an enterprise rated drive so should also be plenty capable. It’s running on ext4 filesystem not BTRFS (I read about many problems using that so I made sure to avoid it).
It is off by quite a bit every single month but everything else seems perfectly fine as far as I can tell (other than the reporting not being accurate of course). If no one else is seeing any major discrepancies I’ll just ignore it I suppose.
8 posts were merged into an existing topic: 4 TB node has been full the past few months, still only 2.17 TBm storage paid