So, what was up with the 37 minute shift? Did I miss something?
Diff change.
[2018-03-05 14:25:59.664+11] workinfo_add(): DIFF CHANGE: hi=512064 delta=9.42% new=3290605988755.0 prev=3007383866429.7
High yet again coz Bitmain's new miners are coming online in their data centres.
Also, could we add another to-do at the bottom of the backlog to put an indicator on the Shifts page when diff changes and restarts cause a shortened shift (similar to how you indicate blocks)? Thanks
That'd be a long way down the todo list
The shift itself doesn't know why it ended, and there's a list of reasons:
pool restart, diff change, halving, block, stale/reject/orphan
so that would involve a somewhat significant code change and a database change, thus why I've never done it,
and then a whole bunch of different indicators ... that really I doubt most people would care
There's already completed code that's in the live system since the middle of last year, that's just not activated due to needing more stringent testing, that completely redoes the shifts, so that they are always on network block boundaries.
It adds an extra level in the share summarisation, called blocksummaries.
(the posts about that are there somewhere in the thread early last year)
That's the major step needed to distribute the pool properly
(as opposed to the node hack done at the moment in ckpool that gets those 5 minute disconnects on 'rare' occasions)
Once that's done much later in the year, then remind me about it
