I am running on Windows, using 14TB drives. When I get within 1TB of a drive being full, I add a new drive, add to config.yaml, and restart the manager. The manager correctly sees the new drive and starts assigning plots to it. The problem is that the last two plots that were assigned to the almost full drive can't fit on it by the time they are ton, and the manager doesn't know what to do with them. Instead of retargeting them to the new drive and copying there, they just get hung indefinitely. I have to manually kill the last two plots on every drive every time when it fills up. Then I have to manually copy the tmp files to the new drive and rename them.
It seems like there needs to be better logic to calculate how full drives are going to be based on all of the active jobs, and/or there needs to be a process to retarget plots to a drive with free space if they can't fit on the intended drive. I can provide logs if needed, this is a fully repeatable issue, it occurs every single time.
I am running on Windows, using 14TB drives. When I get within 1TB of a drive being full, I add a new drive, add to config.yaml, and restart the manager. The manager correctly sees the new drive and starts assigning plots to it. The problem is that the last two plots that were assigned to the almost full drive can't fit on it by the time they are ton, and the manager doesn't know what to do with them. Instead of retargeting them to the new drive and copying there, they just get hung indefinitely. I have to manually kill the last two plots on every drive every time when it fills up. Then I have to manually copy the tmp files to the new drive and rename them.
It seems like there needs to be better logic to calculate how full drives are going to be based on all of the active jobs, and/or there needs to be a process to retarget plots to a drive with free space if they can't fit on the intended drive. I can provide logs if needed, this is a fully repeatable issue, it occurs every single time.