Extremely high ingress + massive node fault

Node Operators troubleshooting
, ,
10

I understand :slight_smile: but Linux while initially looks scary, you get use to how flexible it is…Maybe one day you will feel like using it, but I agree now is not the time to change operating system.

So If it was me, trying to save that node I would;

  1. priority is to get online score back to 100%, that takes 30 days, so this won’t be quick ! but quicker then re-vetting a node for 6+ months.

  2. stop all defrags on that drive, stop any copies to another disk, they need to wait for online score.

  3. edit your node launch script for docker for that node;

a) set the STORAGE allocated to 1TB (replacing the 5.4TB you currently have), don’t worry the node doesn’t suddenly get rid of all the data, it takes months ! the aim is to stop any more data coming in and overloading the node.

b) disable the piece scan on startup, find the nodes config.yaml it will be in the directory with all the node data, add this to the file and save;

storage2.piece-scan-on-startup: false

  1. now start the node, and leave it alone until the online score has recovered.

  2. once online score above 70%, start the copy again to the new CMR disk

  3. once the node data copy, put the Storage allocation back to 10TB, and comment out the piece scan on startup using # at beginning of line on the new disk.

#Warning : Obviously don’t run the node on the old disk, once the node moved to new disk else it will instantly ban node, only run node in one place, on new disk once all data is copied and you updated all the config.

Hope that helps, If anyone notices glaring issues, point them out :smiley:

CP

2 Likes