Folding Forum

Last week 3 of my 7 GPUS stop downloading. I ended up deleting them then added them then they downloaded. Then today the same 3 slots(not the same GPUS) did they same thing. same WS: 140.163.4.200. they finish and never stop downloading.

Ubuntu 20.

Can you please post the log file so we can understand what happened from the client's end?

I checked and 140.163.4.200 has an uptime of about an hour. Thus, if it restarted, you might have encountered this bug: https://github.com/FoldingAtHome/fah-issues/issues/983

I had this too at 00:40 GMT but worked fine on a re-start this morning.

PantherX wrote:Can you please post the log file so we can understand what happened from the client's end?

I checked and 140.163.4.200 has an uptime of about an hour. Thus, if it restarted, you might have encountered this bug: https://github.com/FoldingAtHome/fah-issues/issues/983

Would it be better to put the works servers on a LB?

chabgood wrote:...Would it be better to put the works servers on a LB?

Unfortunately, the Work Servers are owned by various labs and in most cases, are physical machines. Having a physical load-balancer might not be an option for most labs

PantherX wrote:

chabgood wrote:...Would it be better to put the works servers on a LB?

Unfortunately, the Work Servers are owned by various labs and in most cases, are physical machines. Having a physical load-balancer might not be an option for most labs

A Load Balancer points to an IP or DNS entry. I am not sure how that is limiting.

Typically a load balancer is used for connections to multiple servers with identical content or services. That does not apply to the work servers used by F@h. Each WS will have different projects hosted.

What balancing of the load that does exist in F@h is at the Assignment Server level. The AS has information, regularly updated, on what types of projects have WUs available, their priority, and how many. There are also limits set on how many connections can be made for each WS.

oh ok.