Something wrong with client?

[Kebast]

This morning I woke up to all my machines being idle. The last work completed ranged from a few hours to 8 hours ago. One, on the next check for work attempts was in the 4 hour range!
It seems that once my clients fail to get work once, that starts a cycle where they will never get work until I intervene. I paused all machines for one minute, and now every slot immediately pick up a WU and is folding again.
How can I prevent another 4 hour timer on next attempt, and is something else going on with the client?

[jonault]

That's the way the client works. When it can't get a work unit, the time between subsequent requests keeps getting longer. It works well in cases where the client can't connect to the server because it's getting flooded with other requests - slowing down the rate of requests makes it easier for everyone to get through. It doesn't seem to work so well when the server is regularly running out of work units. But that's something that's never really been a problem until the last couple of weeks, so I'm not surprised that the client is handling that in a less that optimal fashion. Maybe the next version of the client will use different strategies in that case. In the meantime, there's not much you can do beyond what you did in this case - pause/unpause.

[BillAnderson]

How about a CPU that hasn't folded anything in almost 48 hours? It's currently sitting at 30 Attempts to get WU. I am so close to giving up on this.