More like Catch22
for days - no work
Now - work available
BUT ya can;t get it 'cause ya can't upload.

I got some! GPUs no longer idle...for now. If the uploads stay stalled I won't be able to download more...

Same here and I guess everywhere..

UL/DL slowly/not possible.. errors..
Also not possible to reach the scheduler..

---

More like Catch22

A Catch two-twenty-two, as all your wingmen are also in the same situation, so even stuff you had pending can't be credited as they can't upload either ;-)

I do wish the User total chart inside Boinc graphed based on the timestamp I completed the WU. I consistenly crunch the same small amount of work everyday. But my chart looks like a I go dark for days on end as my silly wingmen seem to use their systems for other stuff, thus I wait and my graph goes flat ;-)

Finally started to receive work. Actually, at the exact moment when last GPU unit was being crunched. Am I lucky or what :)

---

Yes, you are right, exact description of the problem. upload pending :-(

As others have noted already, bandwidth is pretty much pegged:

http://fragment1.berkeley.edu/newcricket/grapher.cgi?target=/router-interfaces/inr-250/gigabitethernet2_3&ranges=d%3Aw&view=Octets

It's a 100 megabit link at 92 megabits. When you start loading above 80% or so, there are lots of packets that get dropped due to congestion.

It would be nice if BOINC could "flow control" uploads: slowing down all of the attempts would reduce collisions and increase throughput.

In the meantime, hang on and enjoy the ride. It always recovers eventually.

---

Now my "kill network for a week then upload everything, rinse lather repeat" plan isn't so stupid, lol.

God I wish I knew what it is like to worry about being on your last 1,000 GPU units.

---

30+ Computers heading our way! Currently at the "Zomg we need to talk to our tech expert at the co-op about this first!!!" stage. 16 Lab machines and 14+ Staff machines each with 2.2Ghz CPUs and 256MB ram. Think they balance? The RAM certainly is bad

Can we learn anything about these outages, where the bandwidth gets pegged and we all "suffer"? Like, what is the maximum time Matt can leave Berkeley before all hell breaks loose? Or, is there something more interesting?

Uploads keep hitting 100% then just go back to retry in....At least they are making progress.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

The problem at this moment (in about 3.5 moments there will be a new problem :p lol jk), is that everybody's computers are able to get work now, so they use massive amounts of bandwidth. Because of this, nobody can upload their tasks to get more. It's a vicious cycle that only a new internet connection can fix, but right now SETI just doesn't have the cash for it.

---

30+ Computers heading our way! Currently at the "Zomg we need to talk to our tech expert at the co-op about this first!!!" stage. 16 Lab machines and 14+ Staff machines each with 2.2Ghz CPUs and 256MB ram. Think they balance? The RAM certainly is bad

Think it's bad here?

Pigeon transfers data faster than South Africa's Telkom

JOHANNESBURG (Reuters) – A South African information technology company on Wednesday proved it was faster for them to transmit data with a carrier pigeon than to send it using Telkom , the country's leading internet service provider.

Internet speed and connectivity in Africa's largest economy are poor because of a bandwidth shortage. It is also expensive.

Local news agency SAPA reported the 11-month-old pigeon, Winston, took one hour and eight minutes to fly the 80 km (50 miles) from Unlimited IT's offices near Pietermaritzburg to the coastal city of Durban with a data card was strapped to his leg.

Including downloading, the transfer took two hours, six minutes and 57 seconds -- the time it took for only four percent of the data to be transferred using a Telkom line.

SAPA said Unlimited IT performed the stunt after becoming frustrated with slow internet transmission times.

The company has 11 call-centers around the country and regularly sends data to its other branches.

Telkom could not immediately be reached for comment.

Internet speed is expected to improve once a new 17,000 km underwater fiber optic cable linking southern and East Africa to other networks becomes operational before South Africa hosts the soccer World Cup next year.

Local service providers are currently negotiating deals for more bandwidth.

---

Calm Chaos Forum...Join Calm Chaos Now

Can we learn anything about these outages, where the bandwidth gets pegged and we all "suffer"? Like, what is the maximum time Matt can leave Berkeley before all hell breaks loose? Or, is there something more interesting?

If by "we" you mean forum members, I think there are a couple of interesting lessons.

Probably the most interesting is: "when you crush the network, it does eventually recover."

If by "we" you mean SETI@Home, I think there are some experiments that could be run, like setting the "back-off" time to more than 11 seconds.

If by "we" you mean the BOINC developers, I think we have a testbed for some kind of flow control, but someone has to write the code.

---

Think it's bad here?

Pigeon transfers data faster than South Africa's Telkom

Which makes me wonder why they haven't implemented IPoAC in South Africa.

---

Think it's bad here?

Pigeon transfers data faster than South Africa's Telkom

Which makes me wonder why they haven't implemented IPoAC in South Africa.

Thinking about it a little more, I wonder if IPoAC would be possible to improve the bandwidth situation on campus....

---

Think it's bad here?

Pigeon transfers data faster than South Africa's Telkom

Which makes me wonder why they haven't implemented IPoAC in South Africa.

Thinking about it a little more, I wonder if IPoAC would be possible to improve the bandwidth situation on campus....

LOL - but it would take an entire flock!!!!

---

Calm Chaos Forum...Join Calm Chaos Now

Can we learn anything about these outages, where the bandwidth gets pegged and we all "suffer"? Like, what is the maximum time Matt can leave Berkeley before all hell breaks loose? Or, is there something more interesting?

If by "we" you mean forum members, I think there are a couple of interesting lessons.

Probably the most interesting is: "when you crush the network, it does eventually recover."

If by "we" you mean SETI@Home, I think there are some experiments that could be run, like setting the "back-off" time to more than 11 seconds.

If by "we" you mean the BOINC developers, I think we have a testbed for some kind of flow control, but someone has to write the code.

Ok, you have some good thoughts. I would love to know if anyone relevant has read them.

Telkom could not immediately be reached for comment.

I wonder why not?

Telkom could not immediately be reached for comment.

I wonder why not?

Lack of bandwidth...

F.

---

Telkom could not immediately be reached for comment.

I wonder why not?

Lack of bandwidth...

Or lack of pigeons...

Can we learn anything about these outages, where the bandwidth gets pegged and we all "suffer"? Like, what is the maximum time Matt can leave Berkeley before all hell breaks loose? Or, is there something more interesting?

If by "we" you mean forum members, I think there are a couple of interesting lessons.

Probably the most interesting is: "when you crush the network, it does eventually recover."

If by "we" you mean SETI@Home, I think there are some experiments that could be run, like setting the "back-off" time to more than 11 seconds.

If by "we" you mean the BOINC developers, I think we have a testbed for some kind of flow control, but someone has to write the code.

Ok, you have some good thoughts. I would love to know if anyone relevant has read them.

I assume you aren't saying that those of use here are not relevant. :-)

Since every time this happens it seems to be a minor tempest in the forums, I would say that some forum members still haven't figured out that it isn't as big a problem as it might.

For the folks at SETI@Home, I'm sure they've experienced what happens when you try to speed up the recovery process -- my experience is that it generally makes recovery slower.

For the BOINC developers, one of my suggestions is in 6.6.38 and later.

---

Telkom could not immediately be reached for comment.

I wonder why not?

Lack of bandwidth...

Or lack of pigeons...

Same thing.

---