Allow user to have cache and have the user pick time to return results. Spread the times out. When you have a problem and the servers come back on line everyone doesn't connect at same time and crash the server again.
People have extra units from their favorite project so they aren't going to be as upset because they have work to do.

---

Isn't that censoring? The Dutch, Germans, French etc. etc. can only send back results between 12 and 1pm GMT, while the English can only do it between 11am and 12pm GMT. ETC. :)

I think it'll bring on a bigger strain on the server, as it'll have to differentiate between what countries it wants to let in at certain times.

Okay, you said, let the user say when to do it... Say that I have 20 units waiting for upload, I set it to do so at 5am local time, but Joe the American and his whole gang of lovely brothers has 240,000 units to return and he set it to upload at that same time. Do you still think I will get through? I doubt it. ;)

---

Isn't that censoring? The Dutch, Germans, French etc. etc. can only send back results between 12 and 1pm GMT, while the English can only do it between 11am and 12pm GMT. ETC. :)

I think it'll bring on a bigger strain on the server, as it'll have to differentiate between what countries it wants to let in at certain times.

Okay, you said, let the user say when to do it... Say that I have 20 units waiting for upload, I set it to do so at 5am local time, but Joe the American and his whole gang of lovely brothers has 240,000 units to return and he set it to upload at that same time. Do you still think I will get through? I doubt it. ;)

Let's say there is 240000 people doing work. The server goes down and later comes back up. There is a high chance that 240000 people will be trying to connect at once in worse case scenario. Now if it were broken up into blocks of 24 only 10000 would be likely to be connecting at same time. It has helped Seticlassic. It is worked out through Setiqueue though. Seticlassic doesn't have as bad downtime as Boinc setiathome.

---

That would only work if you disabled the Seti@Classic server that day, else they will be down as well. As they have been for the past couple of days. So you have S@C and S@B now uploading work simultaniously, both using about half of the bandwidth they got appointed.

Yet by saying that SetiQueue could do it, it's not strictly said that Berkeley ever done something as stringent as that. And even then, it was a choice to do. How many people did you track down that actually followed it?

Last thing, maybe you want to read around. The technical news and such. When hardware fails repeatedly in your computer, does your little brother blame you personally then, or does he point his finger at the PC you have standing there?

---

That would only work if you disabled the Seti@Classic server that day, else they will be down as well. As they have been for the past couple of days. So you have S@C and S@B now uploading work simultaniously, both using about half of the bandwidth they got appointed.

Yet by saying that SetiQueue could do it, it's not strictly said that Berkeley ever done something as stringent as that. And even then, it was a choice to do. How many people did you track down that actually followed it?

Last thing, maybe you want to read around. The technical news and such. When hardware fails repeatedly in your computer, does your little brother blame you personally then, or does he point his finger at the PC you have standing there?

Look I can see the way work units respond with seticlassic and setiboinc. Since they are forcing people to have a really small amount of work and depend on the main servers boinc servers are down much longer than for seticlassic. The servers can't handle all the users connecting at once.

---

How do you know? We've never had all of the servers and bandwidth available at once for Seti@Boinc!

If you'd read up on the last parts of the technical news you'd be a little bit more informed. SC and SB both look for alien life and are running through the same internet connection and that's where all comparisons between the projects stop.

I know it's difficult to understand, but maybe if you read up a little on threads here and there and on the actual site, then you would be getting more knowledge and wouldn't continue on this old path. :)

For as for the actual cache settings versus benchmarks versus time to connect, they are already under deep review and being worked on. So it's for you to find out what where when, for once.

And please stop using the "reply to this post" button, then copy my whole text along! I know what I said. Just use the scroll while you hold down a shift button until all of the quoted text is gone and you can hit delete.

---

Allow user to have cache and have the user pick time to return results. Spread the times out. When you have a problem and the servers come back on line everyone doesn't connect at same time and crash the server again.

In BOINC there is an automatic mechanism called "Exponential Back-Off" and it does this very thing. When you get a fail on connect, the system starts to "back-off" the time till the next connect.

Where this fails is that some participants will try to use the Update feature to squeeze in and basically negate this feature. The last couple outages for SETI@Home did not cause a single problem, I did not have to hit update, I just let them run and they took care of business. All without me lifting a finger.

And I am running versions from 4.25 to 4.45 and truth be told, I really don't see 4.45 working worse than 4.25 ... a little different, but in general, it works and I don't have to do a thing ...

"Rah! Rah!" ...

Couldn't resist ... Jord sent me an e-mail that put me into "giddy" mode ...

---

Where this fails is that some participants will try to use the Update feature to squeeze in and basically negate this feature.

As far as I know, 4.35 - 4.45 ignore this option. All you do is change the project you're running, while if you have uploads waiting, it may send a quick ping to the server, then ignore it and continue with the countdown.

"Rah! Rah!" ...

Couldn't resist ... Jord sent me an e-mail that put me into "giddy" mode ...

I am glad to have helped. :)

---

And I am running versions from 4.25 to 4.45 and truth be told, I really don't see 4.45 working worse than 4.25 ... a little different, but in general, it works and I don't have to do a thing ...

I have to agree. If I were to keep hitting the update button I'd be one busy person. I simply keep a watchful eye on my RAC and if it's going up I know everyone is happy.

I just loaded BOINC and SETI on a FreeBSD System. I got the boincmgr on my latop watching the progress of the SETI WU on the FreeBSD system.

Both my laptop and PC are running 4.45. All my Solaris systems are running 4.19 and I will be upgrading one of them soon because it is now offically involved in Alpha testing.

Whether you are switching from Classic to BOINC or from 4.19/4.25 to 4.4x it is new and will take some time to adjust. ksnash how about you chill your jets and stop bitching. If you think you know so much why not become a part of the solution instead of continously whinning here.. You and azwoody make a good pair.

---

I'd rather speak my mind because it hurts too much to bite my tongue.

American Spirit BBQ Proudly Serving those that courageously defend freedom.

You and ******* make a good pair.

Now now, I just got Paul into giddy mode... you don't want to have him go down again, do you, CW? ;)

So don't mention the other person. :)

---

HAHAHAHAHA Now I am giddy... ohhh let me delete that quote before you yell at me..

---

I'd rather speak my mind because it hurts too much to bite my tongue.

American Spirit BBQ Proudly Serving those that courageously defend freedom.

Why would the update function be disabled? The number of users that would be online at once is probably small compared to all the active users.

How is the dial-up user going to get his updates, if he can't use the update function, he will need to be hours on line?

After an outage the applications use a random time to reconnect to the server. So when the servers come back online they are not hit form all the hosts at once. The reconnects can be spread out over hours.

---

Red Bull Air Racing

Gas price by zip at Seti

Why would the update function be disabled? The number of users that would be online at once is probably small compared to all the active users.

How is the dial-up user going to get his updates, if he can't use the update function, he will need to be hours on line?

After an outage the applications use a random time to reconnect to the server. So when the servers come back online they are not hit form all the hosts at once. The reconnects can be spread out over hours.

There was a bug for a while. I believe that it was fixed in 4.44, but I know it has been fixed.

---

BOINC WIKI

"Rah! Rah!" ...

??
didn't found a wiki entry.

Ra would fit.

Ra means pharaoh

Once of the most celebres pharaoh called ra (just ra) was wearing the sun disk on his head.

"Ra! Ra!" ...

;)

"Rah! Rah!" ...

??
didn't found a Wiki entry.

Ra would fit.

See: http://en.wikipedia.org/wiki/Ra

Not in mine as I am not documenting Egypt.

---