I started a new passive node and imported the db from a previous node I had. While it is validating the chunks at some point it deletes the whole db (43 gb) and starts from scratch. Any ideas? Could it be permissions or some chunks are corrupted? It definitely doesn’t sound great to delete the whole db without asking. It would be preferable to point at which chunks are corrupted so one can replace them.
impossible, are u sure u copied on the right folder (inside db folder)?
There is the default path for cnode… in case the files are not there it will start from 0
cheers,
Hi Alex, thanks for the quick reply. I always admire how you take the time to help the community with the technical problems.
I have been operating my pool more than a year. I have done this before. Save the db on a usb stick and replaced it so a new node doesn’t start from scratch… This is so weird, I’m suspecting a bug. I tried it multiple times. Steps:
- sudo systemctl stop cnode
- At /opt/cardano/cnode delete db and guild-db.
- Copy the whole saved db folder (43Gb) to /opt/cardano/cnode
- sudo systemctl restart cnode
After a while, when it is trying to validate the chunks, it deletes the whole db and starts from 4.8GB if I recall (8% synced on gLiveView)
I am forced to let it sync from scratch which takes 5 days. Any ideas are welcome. Thanks
Do the old and the new pool differ in versions. I recall that they changed the database layout from 1.32 to 1.33. I read somewhere that it should work by converting and not deleting and downloading again, but maybe it has something to do with that, anyway?