Memory leak present

News updates about the Prohashing pool
User avatar
Steve Sokolowski
Posts: 3169
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA
Contact:

Memory leak present

Postby Steve Sokolowski » Fri Dec 08, 2017 4:51 pm

We discovered a memory leak in the mining servers that is causing them to crash every few hours. When the mining servers crash, the system automatically restarts them, resulting in the affected miners seeing "no valid shares yet" messages. Previously, the load balancing algorithm would then leave that server at 25% capacity and overload the others, but we fixed that.

The problem is that the number of blocks in memory keeps increasing over time, as if some of them aren't being garbage collected due to excess references. I'm investigating the problem and hope to have more information on it tomorrow, although memory leaks are very difficult to reproduce and resolve. Until then, you might see disconnects every six hours or so followed by one minute of downtime while a particular server restarts.
Last edited by Steve Sokolowski on Sat Dec 09, 2017 7:05 pm, edited 1 time in total.
User avatar
Steve Sokolowski
Posts: 3169
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA
Contact:

Re: Memory leak present

Postby Steve Sokolowski » Sat Dec 09, 2017 7:03 pm

The disconnects are still occurring because of this memory leak. Each mining server runs until it uses too much memory and then restarts.

I spent the day trying to configure up a debugger and failed. I'll try again tomorrow, and suspect that it won't take long to figure out the problem after I am able to get it set up. As with anything dealing with computers nowadays, configuration takes longer than coding.

Who is online

Users browsing this forum: No registered users and 3 guests