Hold on I mixed the dates. Today is already the 19th. Let me check.
Edit: Unfortunately my earlier log got wiped in the meantime. So I can only tell from what I had posted. docker ps says the node is up 30 hours now that would be the 04:00Z time from the log. So it seems that “something” caused a node restart. Then it would be expected that the filewalker restarts as well.
The context canceled error shows it was a due to slow disk. So the question is what happened exactly after 24 hours of filewalking that caused this. But there is nothing else on this machine that could affect this nodes storage location. And it would be also a strange coincidence that a process independent from the node software kicks in exactly at the same hour when the filewalker had been started 24 hours before.
In the meantime the node was able to finish the filewalker for the satellite successfully:
2024-08-19T09:41:22Z INFO pieces used-space-filewalker completed
So maybe this is now thanks to the new badger cache that I had enabled for this node, I can’t tell. It would be great if this cache helps to complete the filewalker successfully.
The cache carries over between restarts, but the filewalker does not. So you might not finish a filewalker before it restarts, but all the caching is kept, so its faster after the restart regardless of if the filewalker finished or not. So even if youve never finished a filewalker on a node, with caching enabled, it IS getting faster.
Exactly, and can help to finally to finish it, like in case of @jammerdan. Their node is known that it’s struggling with a filewalker for several versions already, and now at least one is successfully finished. So, with each restart it would be close and close to the end. Not a resume feature, but it’s very similar to it.
Wow, this is the way it actualy works? That’s awsome! I was skeptic enabling the badger, but in this case I will switch it on and enable the start-up walkers on all my nodes. Thanks for this tip!
Yep kinda cool. And AFAIK if any issue arises, you can turn it off and there SHOULDNT be a problem anymore. Might need to delete the cache manually if you want to use it again.