[Passed] Liquidbits.net trustless liquidity provision on USD, EUR and BTC pairs

Currently on ccedk @Cybnate’s LB1.0 works for me but @woolly_sammoth’s 1.21 doesn;t. It has been the opposite before.

I can confirm that. Definitely looks like a change on the CCEDK API.
At the moment I have both bots running - one has connection to CCEDK, the other doesn’t.
I hope that as long as they only change this particular API setting back and forth that at least one of the bot is being able to connect…
Sounds like a desperate workaround :wink:

Hello, dear NU-community. My name is Ivan, i’m developer of CCEDK exchange. I will process any technical issues, related with our api, right here.
Master, could u give me details about error, u’ve meeted while bots were running?
p.s. We have our own bots always runned on ccedk, and each time, when errors happen, we trying to take measures immediately. But for now, our bots are running successfully. So, for exploring possible problems, i have to get more specific information.
Best Regards, Ivan.
p.s. You also may write me directly on email tyax@yandex.com.

1 Like

Hello Ivan,
thank you very much for getting in direct contact with the community!

I have release 1.0 and 1.11 of https://github.com/Cybnate/nu-pool running.
Release 1.0 creates repeatedly errors

could not retrieve ccedk ids, server is unreachable

Release 1.11 is running fine.

That error is generated by the pyBot rather than the ccedk api. Ccedk uses currency ids rather than currency names or codes. The first thing that the bot tries to do is gather these ids from the api.
Until Thursday it was matching on lower case currency codes. Thursday saw the api change to upper case currency codes. Yesterday, I pushed a fix to switch back to lower case. If it’s still causing issues, is it possible that different api servers are returning different cases? (just a guess, I have no knowledge of the api server architecture)

Hello, Woolly. We will follow strict rules policy about API, from this moment, for excluding confused situations, like it was on last week. So, i do not see the necessity in 2nd API server. We sorry for probably inconveniences - we changed api libraries, for getting best efficiency, and didn’t considered a couple of inner differences. From now, api engine will work, stably.

2 Likes

Thanks Ivan, your outreach is appreciated. At least we have some stability right now in order to flesh out any bugs we might still have with our own software. Currently running against your API so that looks good for now.

Maybe creating a “CCEDK watcher thread” would prevent this one from being spammed with messages like this. Nevertheless it might be useful to help people troubleshoot issues.
I mean, why would you want to debug your bot if find out that CCEDK is down?

2015/05/28-07:54:08 ERROR: unable to retrieve order book for usd on ccedk: exception caught: The read operation timed out
2015/05/28-07:54:32 ERROR: unable to retrieve order book for usd on ccedk: exception caught: HTTP Error 502: Bad Gateway

CCEDK appears to be down, because this message " 502: Bad Gateway" is not only generated by pybot, but when using a browser (on a different internet access) as well.

edit: up again

2015/05/28-07:59:47 INFO: ccedk - balance: ...

Hello, master.
From time to time, we update platform and server may return 5xx errors. Usually, these actions prevent working of exchange for not more than 1-3 minutes.
In closest future, we will review our api architecture, for excluding simular situation and running api services in way, not related with exchange engine.
Best Regards.

3 Likes

Hi @dev, thank your for your reply.
I thought about deleting my post when I realized that CCEDK was up again after 5 minutes, but then I thought I might leave it if only to inform visitors of this thread that outages like this might happen and it’s no big deal.

I’m pleased to say that we are now seeing stability on CCEDK API while Bitcoin.co.id is still fine.
There is still a bug wich split up an order in 1 NBT pieces which hasn’t been solved but only rarely occurs.

In the next 24h or so I’m intending to restart the server which will reset all interest counters to zero. Therefore I will do a manual payout after the restart. The daily payout time will also change but still a 24h cycle.The clients should automatically restart. The restart wouldn’t take more than a few minutes.

The restart will synchronise the server with the latest version of the nu-pool software as published by Woollysammoth.

2 Likes

So I can now start considering contributing to nbt/usd on ccedk once again?

If you accept the known 1NBT order bug for now, you can. Woolly is still working on that, but it proves a tough one.
Maybe best to wait after the server restart anyway. Still preparing that. Need some sleep first :wink:

1 Like

Ok tks for the piece of advice and have a good sleep then. :smile:

Just curious, does manual payout not use the interest counter?

Yes and no, we are using a helper program to collect the interest counters from the logs. Based on that output I will send out the pay-outs.

Server restarted a few minutes ago. Will process manual payout in next hour or so.
Just tested the NBT/EUR pair which failed for me before the restart but is now working again! So all good to go.

At the time of writing there are still spaces available for the 0.33%/day rate on both NBT/EUR and NBT/USD pair on CCEDK. It appears to be difficult/expensive to get to the buy side on Bitcoin.co.id (at 0.38%/day) most likely due to the tight spreads and liquidity provided by other bots. Will keep an eye on this and see whether this situation needs review.

server seems to be down

Just noticed. The server process went unresponsive. The server.py process was still running though. The log shows:

2015/06/01-06:24:22 ERROR: exception caught in main loop: NuRPC instance has no attribute ‘txfee’

Haven’t seen this before. Needs investigation.

Meanwhile I’ve restarted the process and it seems to work now. Will prepare another manual payout.

Edit: manual payout sent.

I can confirm that the bot runs on liquidbits.
In case this helps, here’s an excerpt from the log of my bot with comments (time zone UTC)

2015/05/31-04:22:21 server down
2015/05/31-04:25:55 server up
2015/06/01-04:00:29 server down
2015/06/01-04:02:33 server up
2015/06/01-04:24:22 server down
2015/06/01-04:31:51 manual shutdown of bot; waiting for server announced to be up
2015/06/01-06:20:32 manual start of bot, running well