Yep, the mean part in this problem is that BOINC can go into a download loop with several retries per second(!) which keeps it very busy and it cannot recover from this loop without user interaction. This can knockout a lot of PCs that run unattended
Yes, we need to impress on people that they should abort any h1_xxxx (note small h) style work units as the files have been cancelled on the server and even if they crunch and report they will get no credit and their cpu cycles will have been wasted. I've just done this on about a dozen boxes and some were 90% finished when I aborted them. Some hadn't even started so all in all I've saved myself quite a bit of potentially wasted time.
ATTN: Something is seriously wrong with new units
)
Yep, the mean part in this problem is that BOINC can go into a download loop with several retries per second(!) which keeps it very busy and it cannot recover from this loop without user interaction. This can knockout a lot of PCs that run unattended
Yes, we need to impress on
)
Yes, we need to impress on people that they should abort any h1_xxxx (note small h) style work units as the files have been cancelled on the server and even if they crunch and report they will get no credit and their cpu cycles will have been wasted. I've just done this on about a dozen boxes and some were 90% finished when I aborted them. Some hadn't even started so all in all I've saved myself quite a bit of potentially wasted time.
Cheers,
Gary.