I am running storj on a linux box, outside of docker. Node ran fine for about a month then traffic went to zero for several days. I restarted the node and now it is complaining of many pieces missing and the audit scores are decreasing rapidly on all sats. Its already DQ on one and will shortly be on all others. The node is basically trashed. Any idea how this may have occurred and how to prevent it in the future. I will have to discard this node and build it again from zero
maybe post complete system specs, logs
and did you check the drive? (smart; errors)
Which logs would you like to see, syslog is huge
If I grep a missing file in the logs I can see it come in and be referenced a number of times then suddenly its missing. I dont know how to convert the file name into the actual Linux file in order to check it on the drive however
DQ on all nodes now. Node is dead. Time to reformat and try again
Don’t. You need to understand what happened first. Otherwise it will happen again. Reformatting will destroy all evidence.
I would start with checking permissions, mount points, and related stuff.
1 Like
This page shows the blobs folder for each satellite: Satellite info (Address, ID, Blobs folder, Hex)
To convert a piece ID from your logs to a file name: The first 2 characters of the piece ID tells you the prefix folder it’s stored in. The rest of the piece ID is the file name.
An example for a piece in the EU1 blobs folder.
1 Like