I will be doing some main server updates shortly.
First will be an update for ckdb.
As usual the 1st ckdb update will not affect miners, but just the web site will be showing not much useful for the ~15 minutes after I restart ckdb.
The changes involved are all behind the scenes.
I've added some new lock checking and (100%) deadlock prediction code to get things ready for threading some of the threads that can block and slow things down on rare occasions.
ckdb + ckpool is fine at 6PH (as we saw yesterday when we found that block) and all probably would be ok well above 40PH, but there's ways I can make sure of that and make that upper limit higher ... so with this change I'll be able to verify all is OK to go ahead and speed ckdb up even more.
I don't expect any problems with the changes even though they are rather extensive:
(as can be seen here:
https://bitbucket.org/ckolivas/ckpool/commits/c7f03c2a331f111696f0e8163f295461711e8eb5 )
but of course they don't affect ckpool running so the only side effect could be some restarts of ckdb due to new ckdb code problems - that I don't expect that at all.
After that I'll also do a ckpool restart.
There's some changes in there that may help improve performance also - but ckpool is far from having any performance problems on the pool.
For the majority of users, a ckpool restart should just show as a reconnect in cgminer.
Though some minority of users will see a failover since not everyone reconnects fast enough
- inside the 30 seconds required to avoid a failover in cgminer.
I'll post again once it's all done.
Edit: ckdb restart complete - web site all back ok
Edit2: yeah the 2nd step was done also (about 8 minutes ago) ... now almost back to where we were before that