Closed kuzdogan closed 1 year ago
After stopping both IPFS nodes, now the server has at 60-80% of the full CPU capacity. Before it was always at > than 100%.
Pinned an up-to-date version of the repository on a local node and published under the IPNS https://ipfs.io/ipns/repo.sourcify.dev/
It wasn't possible to directly import a CID to pin in Infura, and their ipfs-copy
didn't work well. For web3.storage, waiting for approval to use the bulk pinning API.
Closing it for now as local pinning seems to be sufficient for now. Will check back next week.
As we are facing resource overload, we need to turn off the IPFS node for some time. For that we'll temporarily outsource the pinning externally. The downside will be, we won't be able to update the IPNS for a while with newly verified contracts, but we will pin the newly verified contracts separately.
Preliminary
Do these before starting the rest:
1. Persist the data
.ipfs
folder and do not necessarily do a completeipfs add
when the ipfs container starts running. This is actually long overdue, and will make sure we can pick up where we left, in case something goes wrong in the next steps, and have to turn on the ipfs-server again.2. Check performance increase
Server
.ipfs
folder before shutting down the container, or is there a way to persist it on the go. I.e. will those files vanish if we change theipfs.yaml
and restart the container?Migrating the pinning
ipfs add
viascp
etc. such that the process will run on another host but the files will be pulled from the server machine. This will not consume extra resources on the server and we don't need to copy everything.ipfs add
with external pinningipfs pin remote add --service=estuary $rootHash --background --name=$pinName
Migrating the IPNS
Changing how we pin in the server
Finally
Gateway
View in Huly HI-447