Node failed to calculate data changes online

Somewhere is error in space calculation.


This node is 218h in running in row
and below is after restart, no configs was changed just restart.


looks like it forget about uncompacted garbage at all and not stopp upload new data till it hitt disk full.

Yes, uncompacted garbage is not included in the chart (it shows only the space occupied by pieces). I agree: it’s super confusing. This will be changed very soon: Fix UsedSpace calculation of Storagenode and Multinode web console · Issue #7700 · storj/storj · GitHub

It’s a display only issue, so it doesn’t affect the free space reporting. Still Satellite should stop sending new pieces if there is no more free space.

If it doesn’t work it’s a different issue, and we are planning to investigate it. Do you use dedicated free space calculation by any chance?

Yes, it not stop ingress till hdd is full.
I dont use dedicated, because it not show stats then.
It overused about 400GB so it not only visual bug.
when node start it looks like it read it all as should do it, but after some time it go totally wrong.

1 Like

Also, i found that used_space_byprefix is also make node show wrong data, but to bigger size, so i usually delete it when i have possibility.
I have only hashstore data on every node

So main question, How can I help?