Status as of Saturday, October 28, 2017

Good morning!

• Chris discovered that one of servers closer to the edge of the network also had a connection limit. In this case, the limit affected all servers because connections passed through that system onto the rest of the network. The cause is that Debian 9 uses nftables, not iptables as older versions of Debian did. We are using Debian 9 on all new installations now, so this is the first time we encountered the problem. He fixed it this morning, which caused some downtime during the fixes, but everything is back to normal now and there should be fewer network connectivity issues.
• Today, as promised, we will be spending the weekend focused on resolving all old support tickets. We are right now up to October 10, although there are some customers who may have replied to tickets asking for an update and that pushed the "latest reply" date to the future. Regardless, we hope to reply to all of the tickets this weekend.
• Also this weekend, I think we figured out a way to deal with the "forgot password" tickets permanently. More details will become available when we complete the implementation.
• There are still just a few too many issues right now for us to feel comfortable expanding capacity. While performance is still acceptable for now, if it does become an issue, the system has become big enough that I don't think we should expand our services until we have put additional work into the issues that generate tickets and make the system less stable (like the password resets and the connections limit we just fixed.)

Hiya,

I still have several miners that wont connect. tried everything.
solo not been working for any coins I have tried for a week or so.
a=x11 n=x9 i.e. having probelms connecting. its been pretty rare to have them all connected together. I don't want to mine anywhere else

Any ideas? anything I can try?

Thanks for help!
Phil D.

Efficiency today is really poor, 85-87%

The system is approaching capacity. There is one more trick I think Chris can try, but we're going to close registrations soon because efficiency will go down as shares can no longer be processed in real-time.

With efficiency already down for a lot of users 10%, is the pool already at that stage where they aren't processing in real-time now?
Are we running on swapper/virtual/disk cached system memory now that would already be affecting the efficiency?

Steve Sokolowski wrote:The system is approaching capacity. There is one more trick I think Chris can try, but we're going to close registrations soon because efficiency will go down as shares can no longer be processed in real-time.

Hey Steve & Chris, which capacity of the system is near limit(s) :... memory, CPU, disk space, network connection(s) ...what exactly is at capacity ?

I am a computer engineer too, Master degree in Computer Science, 35+ years experience ... maybe could help at least in brainstorming or something ?

You could PM me if this is out of the scope of this topic.

Regards,

ZiGen

ZiGen wrote:

Steve Sokolowski wrote:The system is approaching capacity. There is one more trick I think Chris can try, but we're going to close registrations soon because efficiency will go down as shares can no longer be processed in real-time.

Hey Steve & Chris, which capacity of the system is near limit(s) :... memory, CPU, disk space, network connection(s) ...what exactly is at capacity ?

I am a computer engineer too, Master degree in Computer Science, 35+ years experience ... maybe could help at least in brainstorming or something ?

You could PM me if this is out of the scope of this topic.

Regards,

ZiGen

Hi Zi,

Thanks for the help. Unfortunately, development isn't the type of assistance we require to fix this problem. The code is already implemented, but not tested. We need people to deal with the financial side of the business so that we can decide whether to make an investment in the mining equipment required for testing. If we do need development expertise in the future, we'll be sure to contact you.

Thanks,

-Steve