Bug in new Vetted Date

Thanks for the reminder. Fixed that now!

4 Likes

Congrats on repairing 11 nodes with a relatively untested workaround/hack but I was busy doing other things.

Myself and a few others in this thread have been talking about this for quite a while and the issue was ignored/denied. Now it’s been proven that there’s an issue, I’d prefer an official fix rather than suspending a bunch of nodes and hoping they come back.

I have 17 nodes of which 8 are showing unvetted Satellites. Some are showing unvetted on 2 Sat’s and it always seems to be EU1 and/or AP1. US1 and Saltlake are fine.

The one above is the youngest node.

1 Like

I don’t think that last one was directed at me, but just in case. Before Storj added the satellite’s report on vetting status to the nodes, I had no other way to determine vetting status in the earnings calculator then to assume 100 successful audits meant it’s vetted. So I couldn’t display that information, because the node simply wasn’t aware of it. This only recently became available to the nodes.

And it wasn’t in master, because I asked some users to verify it was working correctly, since I didn’t have any nodes in vetting nor nodes showing this issue. I didn’t want to break people’s scripts who update automatically when I release a new version. There’s been sufficient testing now though, so I updated the master version.

No, wasn’t aimed at you.

It was referring to a thread I started a long time ago. A few others jumped on too and experienced a few nodes receiving considerably less traffic on the EU1 satellite despite identical configs/hardware. It was dismissed as ā€œone of those thingsā€ and ā€œluckā€. As you pointed out - none of us had visibility of what was happening on the Sat’s so we hade to take the dev’s word as gospel. And now, here we are, their update has allowed your script to show… Oh, an actual issue.

As for the 610 posts in, not really you either, just rather your post should be a pinned and easy to find paged under the Node Operators. Trying to find ā€œUseful Toolsā€ (yours being a very useful tool) on this site is a nightmare if you don’t know exactly where to look. The tools should have their own section with a dedicated menu plastered across the top, not buried under a tag on the node ops forums.

4 Likes

Afternoon all

As one of the original posters here, what was the conclusion on a plan/solution here? Thinking about anyone else who drops by this thread too - im not exactly clear I wish to get the node suspended to ā€œtryā€ and fix or should we give it a try? Hoping, if fixed, I may get some local EU traffic

Many thanks

CC

Hi @Climbingkid , thank you for highlighting the need for more clarity on this issue.
I’ve forwarded your questions internally and someone will return here with answers as soon as possible.

3 Likes

I checked github but found no code changes related to this problem. So even in best case it will take some more weeks to get a fix rolled out. I therefore started to apply the hack mentioned above to all affected nodes.

3 Likes

I’m lazy. If Storj thinks they have too many mistakenly-unvetted nodes: they have all the tools they need on the satellite side to fix it themselves.

1 Like

It’s about 10% of node/satellite pairs affected. So given the current oversupply of nodes no need to fix it at all.

1 Like

I’m lazy too, but I like to tune stuff and try things, especialy if there are added benefits in the end. The software is so mature now and I tuned my nodes to ā€œperfectionā€, so no need to play with them anymore, and I’m kind of missing it… when a rare ocasion like this arises, I’m all in. :blush:
And waiting is not my strong suit.

2 Likes

I run this storj node now for 12 months. The last updated added the vetted status on the dashboard. I see now that I am not vetted on this one? I have also almost no used space on this server.

All other say vetted and I store currently 9TB. Is this normal?

BTW: suspension, audit and online are all 100% on every server.

Thanks for your Reply!

Lots of SNOs are seeing the same thing: thread is here.

1 Like

Thank you - so its a bug (although I feel like it is a bug that should not happen in the first place…).

I will wait it out

2 Likes

There are no bugs in production software. Only some undocumented features.

4 Likes

Hi all,

We have executed a query to align the vetted date that we register in the reputations table (used by audits) and the nodes table.

As some of you commented here or in the Github issue, we may lose updates in the nodes table because the column is updated in both tables but without a transaction. However the locks doesn’t back up the issue.

We will have the issue open until we get the time to work more on it.

Thank you for your collaboration.

15 Likes

Hooray for the great revetting! Un-unvetting? Delayed vetting? Vetting v2? Whatever… :winking_face_with_tongue:

3 Likes

Hello!
Is this normal? Or it is a bug?

They’re applying a workaround for now…

Hiya all

I’ve been quiet with my comments on this matter, since it does not really bother me.

I had a handful of nodes that did not report the correct vetting status - all of them fixed themselfes yesterday.

Good day to you, Storjlings

2 Likes