Hi all,
I used the instructions on migrating to Docker from Windows. After several days of copying with the node still running, did RSYNC also delete the files at the destination that were deleted by node while running? I ran RSYNC several times and the last time no more files were copied and no errors since the node was off.
As always thanks for your replies.
The last command for delete should be with the node stopped.
Out of curiosity, does the node ever audit its dataset and clean up its directories? This particular node has been running close to 2 years and I can’t imagine that there hasn’t been errors over time causing things to be be missed. Its about 9TB so that’s a bunch of data to be perfectly synchronized.
The satellites audit your node by asking for challenges that shows them your node is still storing the files. It doesn’t do individual piece repair though. It just judges whether your node is reliable. If not, you get disqualified. If it is, the network can deal with any minor loss of pieces that may have occurred.
But I don’t agree with the attitude that something must be wrong or out of sync. Unless you made a mistake or hardware is having issues, your node should still have all pieces and never fail a single audit. My three nodes never have.
Well, I was just wondering if there is some “house cleaning” function that Storj has besides looking for data that may be missing. My nodes haven’t failed any audits as well minus the one that had drive issues a good bit ago. But if their system always deletes and adds things without fail, then I guess their is no need to analyze it’s own data. However, they could be robbing themselves of storage space if they have files on the node that were accidently left there for whatever reason.
Again thanks for all you do @BrightSilence and others in the community.
Well there is garbage cleaning, which uses bloom filters that allows the node to remove anything that doesn’t belong there anymore. This is mainly for when nodes were offline when the deletes happened or when repair happened while they were offline. So that takes care of the “files that don’t belong there” part. For files your node should have, your node only sends out proof of storage when the file has been stored on your node. So satellites won’t assume you’ve received it without your nodes confirmation. So at the least you had the piece at one point.