Then it switched to 2 other failovers but they could not connect to any.
I had to close program and start it again. There were no connection/network problems. Tried ping and that worked, other rigs were working without problems.
Any hints on it ?
This is the infamous reconnection issue. We are working on it, and it should be fixed in 2.4. It is maddeningly inconsistent - during the tests a hundred reconnects in a row go without a hitch and then it strikes again.
Is there any option like -minspeed ?
Not yet but we are considering it. However to be truly useful, it should probably have the ability to set a per-GPU minimal speed also if the GPUs are not the same.
Other bug:
Using blockchain drivers, having these results with 2.3:
17547:23:08:20.300: main GPU1: 48C 90%, GPU2: N/A, GPU3: N/A, GPU4: N/A, GPU5: N/A, GPU6: N/A, GPU7: N/A
No difference over 2.2b with fan/temp reporting.
This is interesting, as HW info reporting problems were only observed on the latest Adrenalin drivers up until now. Could you please send us a full log from a test run? To avoid posting your wallet address, you can run
PhoenixMiner.exe -bench
for a minute or so (until you see the first GPU temperature/fan report) and then send us the log file contents (on PM if you want to, we can receive PMs from all members).