Log Message - ERR no delegation distribution info

[cfl0ws]

Background

During the network restart using v10.0.0 our logs displayed only the following message immediately after restarting osmosisd.

ERR no delegation distribution info

Upon reporting this to the Osmosis Validators discord channel, @ValarDragon explained the message was due to "no swaps in that pool" and asked me to create this issue so the log messaging could be improved.

Suggested Design

TBD

Acceptance Criteria

TBD

[SlingshotCade]

I too have run into this problem after upgrading to Osmosisd 10.0.0. Once this issue shows up in the logs then the API stops responding (port 26657). A couple times I restarted cosmovisor and it started to work again. But then the "ERR no delegation distribution info" shows up and then the server stops working. Is there any more information on this?

[ValarDragon]

The API is down generally due to epoch time, which is a more a longstanding problem with queries aren't available during block execution.

[mattverse]

After investigation, noticed that we had this error not because the pool didn't have any swaps in between epochs, it's because we don't delete intermediary accounts. Working on a PR with the fix

[ValarDragon]

Hrmm, don't we need to be careful wrt when we need to delete intermediary accounts? (IIRC theres something in the spec about this)

[mattverse]

Can't find anything in spec as of now, the current plan im thinking of is deleting intermediary accounts that doesn't have any connections & delegations any more, I cant think of any instances where deleting can be dangerous atm.

If you feel concerned about it, we can also simply pass without returning error here?

[effofxprime]

Regarding this issue, since I also face the same error, is it possible that it's causing a performance hit that I've described here in this ticket (#2173) ?

[mattverse]

Epoch has performance hit mainly because of distribution logic and internal logics that's happening in epoch time, shouldn't have relation with the error log we see here

[ValarDragon]

I believe this has been fixed