I creating this topic to pay attention to the Graceful Exit (GE) implementation and node selection on the satellite side during GE. So, Let’s start to review GE process like an engineer. Introduction: I started GE on one of my storagenode for satellite.stefan-benten.de (118UWpMCHzs6CvSgWd9BfFVjw5…

[image] Odmin: @littleskunk So, I think error handler already implemented? I like your changelogs, could you please tell a little bit more about how GE is working? (especially how to remote errors is handling) I have no information, just a few posts from @Alexey and that all… GE node tells t…

as example 2020-07-17T01:09:24.841+0300 INFO piecestore uploaded {"Piece ID": "P3W3LYGPH6K2UWBHKE6V3U3GSSD67VFVR3BLDPZ2QNCYAIAPKP3Q", "Satellite ID": "121RTSDpyNZVcEU84Ticf2L1ntiuUimbWgfATz21tuvgk3vzoA6", "Action": "PUT"} 2020-07-17T01:09:24.842+0300 INFO piecestore upl…

[image] Odmin: I will just remind: This if two way trust, so if you not trust… and every SNO can ask you: Yes, of course, that’s how distributed networks work. Participants don’t trust each other implicitly. Storj Labs has to build up a reputation that it will pay for storage as agreed. If yo…

[image] Krey: "error": "write tcp 192.168.156.4:38344->78.94.240.189:7777 This makes me believe something is unable to handle the connection count or DNS queries. To be fair. Yes the storage node should be able to handle it. I am not 100% sure why this is happening. I am sure it is not a pro…

[image] thepaul: Yes, of course, that’s how distributed networks work. Participants don’t trust each other implicitly. Storj Labs has to build up a reputation that it will pay for storage as agreed. If you feel that this thread is evidence that we don’t pay as agreed, you should remove Tardigrad…

[image] littleskunk: This makes me believe something is unable to handle the connection count or DNS queries We can see “write tcp error” on any node any time. Logs contains ton of it on any single host with single node, my or on cloud dedicated servers. Guided by your example, I see one poi…

[image] thepaul: I’m not trying to diagnose this case. I don’t know anything about the specific case. I was only responding to the claim that Storj Labs disqualifies nodes for errors out of the node’s control. My opinion: Responding and ignoring details it not fair, we here try to find the ro…

[image] Krey: Guided by your example, I see one point of failure here and this is not a server but your software. For a moment I was hoping that we would try to find the root cause. Ok then I have to agree @thepaul . If you are saying it must be on the target node side then I have to say afte…

I’m sorry if I’ve added too much noise. Maybe it can be split out to a separate topic? I don’t know how that works. But I couldn’t let a claim like “Really satellite can and do DQ you node for errors not caused by you storage node.” sit uncontested. The Storj network is not perfect and has room to …

Ok, I pay attention to more robust log analizer and count all tries and pieces.

But does it make sense to do it at all? Stefan is closed, in a couple of days you will clear everything and even if I do an additional analysis, you will not be able to finally sort out my problem? To be clarify, money is not important to me, I am quite happy with payments from benten. I just want …

Storj Community Forum (official)

Graceful Exit (log inspection)

Developer Discussions troubleshooting

show post in topic

Home
Categories
Guidelines
Terms of Service
Privacy Policy

Powered by Discourse, best viewed with JavaScript enabled