Why don't you publish the LTD figures as JM7 suggested. Then we could see exactly what is going on and give you a realistic estimate of when EAH will start again.
You do realise that the scheduler is just obeying your choice of resource share and is simply not running EAH because EAH has already had more than its fair share for the moment, don't you?
If you lower your "connect to network" interval to 0.1 just until things stabilize you will find things will sort themselves out much more quickly. Once that happens, you can then start bumping it up again bit by bit every couple of days using a sequence like 0.1 - 0.5 - 0.75 - 1.0 - 1.25 - 1.5 and watching carefully for any signs of stress until you find a setting that gives you plenty of work and will also round-robin the way you would like with no need for prolonged "droughts" like you currently have. I would hazard a guess that something like 0.75 - 1.0 will turn out just fine and that larger values will continue to throw you into "drought mode" like you are currently experiencing.
Others have suggested lowering the value. It really will work.
Well it seems to request an amount of work based on your processor speed and connection time. For instance, one work unit on my host represents about 7 hours of CPU time. I want a queue length of 2 work units, let's say. I therefore need a connection gap of 7-14 hours, which is 0.3 to 0.6 days.
If you only plan to connect once per week, expect to be sent work units for a week's worth of CPU-time. That will be too much if you are running multiple projects.
Work started to download today, finaly, Einstein has a 6000 sec long term debt.
But you really haven't solved your problem. If you have a look you will see that each machine got 8 work units which is your daily limit. You may even get more for each machine tomorrow, depending on how high you have left your connect to network interval. So as the deadline approaches, once again you will be on the merry-go-round and the deadline dance will start all over again. You will have a couple of days solid EAH with nothing but EAH, followed by a prolongued deadline drought where once again you can give us a daily update of no new EAH .....
You mentioned previously that you were supporting three projects. I don't recall ever seeing the resource share but let's assume you gave EAH one third. This means that, to honour that share, you probably only need one EAH unit per day!! Yet you just allowed your boxes to download 8 each (maybe even more). Can't you see that you will never solve the deadline dance problem that you have created??
Why don't you just do the perfectly simple fix of setting your connect interval to a much more sensible value???
Work started to download today, finaly, Einstein has a 6000 sec long term debt.
But you really haven't solved your problem. If you have a look you will see that each machine got 8 work units which is your daily limit. You may even get more for each machine tomorrow, depending on how high you have left your connect to network interval. So as the deadline approaches, once again you will be on the merry-go-round and the deadline dance will start all over again. You will have a couple of days solid EAH with nothing but EAH, followed by a prolongued deadline drought where once again you can give us a daily update of no new EAH .....
You mentioned previously that you were supporting three projects. I don't recall ever seeing the resource share but let's assume you gave EAH one third. This means that, to honour that share, you probably only need one EAH unit per day!! Yet you just allowed your boxes to download 8 each (maybe even more). Can't you see that you will never solve the deadline dance problem that you have created??
Why don't you just do the perfectly simple fix of setting your connect interval to a much more sensible value???
RE: RE: RE: RE: Quote
)
RE: RE: RE: RE: Quote
)
What do your LTD and your
)
What do your LTD and your STD's look like today?
Also, you said that you threw in one or two more projects, so it could take more than the 7 days that you've set as a deadline.
Jim
Jim
RE: As of Day 6 no
)
Why don't you publish the LTD figures as JM7 suggested. Then we could see exactly what is going on and give you a realistic estimate of when EAH will start again.
You do realise that the scheduler is just obeying your choice of resource share and is simply not running EAH because EAH has already had more than its fair share for the moment, don't you?
If you lower your "connect to network" interval to 0.1 just until things stabilize you will find things will sort themselves out much more quickly. Once that happens, you can then start bumping it up again bit by bit every couple of days using a sequence like 0.1 - 0.5 - 0.75 - 1.0 - 1.25 - 1.5 and watching carefully for any signs of stress until you find a setting that gives you plenty of work and will also round-robin the way you would like with no need for prolonged "droughts" like you currently have. I would hazard a guess that something like 0.75 - 1.0 will turn out just fine and that larger values will continue to throw you into "drought mode" like you are currently experiencing.
Others have suggested lowering the value. It really will work.
Cheers,
Gary.
Well it seems to request an
)
Well it seems to request an amount of work based on your processor speed and connection time. For instance, one work unit on my host represents about 7 hours of CPU time. I want a queue length of 2 work units, let's say. I therefore need a connection gap of 7-14 hours, which is 0.3 to 0.6 days.
If you only plan to connect once per week, expect to be sent work units for a week's worth of CPU-time. That will be too much if you are running multiple projects.
RE: What do your LTD and
)
Work started to download today, finaly, Einstein has a 6000 sec long term debt.
RE: Work started to
)
Glad to see that things have started to stabilize for you.
Jim
RE: RE: Work started to
)
Patience is a virture, Ya right! Ha!
RE: Work started to
)
But you really haven't solved your problem. If you have a look you will see that each machine got 8 work units which is your daily limit. You may even get more for each machine tomorrow, depending on how high you have left your connect to network interval. So as the deadline approaches, once again you will be on the merry-go-round and the deadline dance will start all over again. You will have a couple of days solid EAH with nothing but EAH, followed by a prolongued deadline drought where once again you can give us a daily update of no new EAH .....
You mentioned previously that you were supporting three projects. I don't recall ever seeing the resource share but let's assume you gave EAH one third. This means that, to honour that share, you probably only need one EAH unit per day!! Yet you just allowed your boxes to download 8 each (maybe even more). Can't you see that you will never solve the deadline dance problem that you have created??
Why don't you just do the perfectly simple fix of setting your connect interval to a much more sensible value???
Cheers,
Gary.
RE: RE: Work started to
)
I use dial up!