Status as of Sunday, October 4

Discussion of development releases of Prohashing / Requests for features
Forum rules
The Development forum is for discussion of development releases of Prohashing and for feedback on the site, requests for features, etc.

While we can't promise we will be able to implement every feature request, we will give them each due consideration and do our best with the resources and staffing we have available.

For the full list of PROHASHING forums rules, please visit https://prohashing.com/help/prohashing- ... rms-forums.
Locked
User avatar
Steve Sokolowski
Posts: 4585
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Status as of Sunday, October 4

Post by Steve Sokolowski » Sun Oct 04, 2015 10:47 am

Good morning! We'll be issuing a new mining server release tonight, along with the website and trader later. The release resolves a few issues.
  • The trader is currently very conservative when an unhandled exception occurs during buys and sells. The mining server responds to these exceptions by placing the coin into an excluded state. However, there are times when the issue that caused the exception is temporary, like a one-time disconnect that causes a buy or sell to fail. Now, the exception needs to happen repeatedly before coins will be excluded, increasing profitability by not excluding coins unnecessarily.
  • Currently, some miners get stuck in work restart testing for an unknown reason. We think the causes of getting stuck are miner-specific, and we can't reproduce them. Therefore, we decided to add a new feature where, if a miner does not respond to work restart testing within 50s, the miner is assigned a suboptimal work restart time of zero. That way, the miner will still be able to mine, and users can still bypass the testing if they choose to do so with the r= parameter. We were originally going to address this problem with additional documentation, but decided we should offer suboptimal mining instead so that miners then are incentivized to remain and adjust their settings.
  • Some more minor issues will be addressed with the website. There will be some text changed on the earnings page to more accurately describe what the graphs are displaying. Users can no longer delete bitcoin payout addresses, because if they choose 100% in a coin other than bitcoin, they still need to receive bitcoins when that coin is excluded.
Also, we caused a crash in Cryptsy's markets this morning because Cryptsy's API was timing out on Thursday and Friday, so we held coins to sell until they fixed the problem. Finally, kires said that he can't see the "like" and "g+1" buttons that we added to the forums; however, nobody else is able to reproduce the problem. If you don't see the buttons, please let us know.
dman99
Posts: 7
Joined: Wed Sep 30, 2015 7:24 pm

Re: Status as of Sunday, October 4

Post by dman99 » Tue Oct 06, 2015 1:26 am

My miners were assigned a work restart time of zero. Not sure why they weren't responding to to the work restart testing. I stopped them for a minute and then started up and all seems to be well.
User avatar
kires
Posts: 188
Joined: Thu Mar 05, 2015 8:25 am

Re: Status as of Sunday, October 4

Post by kires » Tue Oct 06, 2015 9:11 am

The same thing happened to me. Bouncing the miners resolved my issue, as well.
User avatar
Steve Sokolowski
Posts: 4585
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Re: Status as of Sunday, October 4

Post by Steve Sokolowski » Tue Oct 06, 2015 8:07 pm

I looked into this issue. It turns out that when the mining server is restarted, some mining servers pause for 60s and then try to submit a share. The work restart testing is configured to cease after 30s, so when the server is restarted, miners that were previously connected can sometimes wait too long to submit a share and the server thinks they are not responding.

I'm going to add a fix so that the server doesn't time anyone out on the work restart testing at initial startup for 300s instead. You won't encounter this problem again until we restart the server for the next release (that will contain this fix), so you can consider this issue effectively fixed.

Thanks for the report!
Locked