Is there any reason not to add skein, the 5th Algo for DGB/MYR/AUR?
I'll be added a few more algos over the next few weeks.
btw, the pool is already running of SSD's and is also distributed over 4 8-12core servers with 256GB ram. It's not the hardware. I've changed it 3 times now. The hardware investment is now 10-15k a year. I'm still looking for a dev that is willing to re-write port the stratum component to something more robust and stable. I have one prospect so far and it by no means cheap. It will be a $2-3k job and has to be done right. Trust me, nothing I hate more than seeing the stratum resets. It's also
BTC out of my pocket too.
My end game is to be A) The most profitable, B) Reliable, C) Stable and D) The biggest.
I've got 3 out of 4 IMO, stability will come as soon as I can get it developed.
Ouch. I hope you get a highly configurable stratum server for that job. It's a bit beyond something I would want to tackle at the moment.
The issue I was getting involved the monitor thread. I was trying to use it to solo mine some MYR which now has the getwork protocol removed. For some reason the system thought my MYR address was not a valid MYR address, so I think I was getting this due to the stratum server not having any jobs to provide.
If your stratums are resetting at the same piece of code (perhaps due to getblocktemplate rpc delays), than I suppose you could always try increasing the max job delay. It might mean the stratum clients aren't being fed with jobs to run at full whack at times, but it might be better than a reset.
stratum.h
#define YAAMP_MAXJOBDELAY (2*60)
stratum.cpp
///////////////////////////////////////////////////////////////////////////////
void *monitor_thread(void *p)
{
while(1)
{
sleep(120);
if(g_last_broadcasted + YAAMP_MAXJOBDELAY < time(NULL))
{
stratumlog("%s dead lock, exiting...\n", g_current_algo->name);
exit(1);
}
}
}
///////////////////////////////////////////////////////////////////////////////
