Is this supposed to have been fixed?
Because I have just checked again on one node with this error message on two randomly selected pieces and again the pieces have been delivered into the correct blobs folders while the uploads have been logged as failed.
This is a very unpleasant situation. As this error message shows up very frequently in the logs making it impossible to judge if the node has an issue or if pieces are successfully uploaded.
Interesting edit: It seems that the node got updated couple of hours ago. I see no a new error message:
INFO piecestore upload canceled (race lost or node shutdown) {"process": "storagenode", "Piece ID": "XXXX"}
However I don’t know if this is helpful. First of all, there is no indication of the satellite any longer in that message. So if I want to check if the file has been uploaded or not, I have to check for it on every satellite now.
And still after I have checked I see the piece has been uploaded.
So basically nothing has been solved and the questions are still the same:
- Why is this considered a failed upload when the piece is present in the correct blobs folder?
- Why does the piece not get outright deleted resp. moved into the trash folder, if it is a failed upload?
- If the upload is considered not successful this means I don’t get paid for the space it occupies, there is no chance for egress on this piece and it is simply wasting space.