Closed ArseniiPetrovich closed 2 years ago
Can you elaborate on why you see this as being an issue @ArseniiPetrovich? It is not a surprise to me that running out of chain disk space would result in chain corruption and maintaining disk space is something that needs to be monitored to avoid. It can also be easily resolved by importing a new lightweight snapshot.
@TippyFlitsUK not so easy for an archival nodes that have all the chain state :) Sure, disk space need to be monitored and it's purely our fault that we overlooked this alert in our systems. However, chain corruption when having a lack of disk space still have to be considered as a bug, at least from my point of view, no matter "surprise" it or not, because it makes even a simple mistake to have great consequences. Can't we verify the available space before writing there or at least deploy a kind of recovery tool that allows you to rollback to several blocks behind the chain and resync?
Thanks for the clarification @ArseniiPetrovich! Agreed that this presents a far bigger problem with archival nodes. I don't agree that represents a bug though. Can you please file a new ticket using the enhancement request form and provide the additional info requested. Many thanks! :pray:
Checklist
Latest release
, or the most recent RC(release canadiate) for the upcoming release or the dev branch(master), or have an issue updating to any of these.Lotus component
Lotus Version
Describe the Bug
Here at Lotus nodes we unfortunatelly run out of disk space recently on one of our archival nodes on calibrationnet. It was running 1.16.0, and when we restarted it failed with the following issue:
I tried to upgrade to 1.17 as suggested at https://github.com/filecoin-project/lotus/issues/8916, but it didn't help. Is there any chance to recover from this condition? Thank you!
Logging Information
Repo Steps
lotus