You know..

Why can't BOINC contact the server to report the 47 LHC (or SETI, or rosetta.. etc..) workunits I completed when they are all finished?

Why does it wait several hours before doing so?

Shouldn't the program start attempting to contact the server to report the 47 completed results once they have all completed?

---

You know..

Why can't BOINC contact the server to report the 47 LHC (or SETI, or rosetta.. etc..) workunits I completed when they are all finished?

Why does it wait several hours before doing so?

Shouldn't the program start attempting to contact the server to report the 47 completed results once they have all completed?

Reporting Work - with credit to John McLeod VII

There are two parts to returning work. First, when a result is completed, it is immediately uploaded and marked as ready to report. Later, it is reported to the scheduler. This is so it can report batches of WU to the scheduler (which is a separate sever than the upload/download servers), and saves on bandwidth for you and the project.

The BOINC reports completed work at the first of:

1) The next work request from a project.
2) 24 hours before the deadline.
3) Immediately if the work is completed later than 24 hours before the deadline.
4) Connect every X days after the work is completed.
5) With the next trickle (CPDN only at this point).
6) When the user clicks update.

---

Boinc V 7.4.36
Win7 i5 3.33G 4GB NVidia 470

Well, I understand all that, but Why can't it also report the completed work when it is finished with all of work from one project? Surely project owners want to have their work as soon as it is finished...?

---

Well, I understand all that, but Why can't it also report the completed work when it is finished with all of work from one project? Surely project owners want to have their work as soon as it is finished...?

Unless you carry a very large cache and have multiple projects, the reporting of finished work will occur well within 24 hr with normal round robin activity. This is plenty fast enough for all projects.
As far as I know, only LHC intentionally run out of work as part of their normal operating procedure and would like result returned quickly. That is, in part, the reason for their short due dates. They need the data analyzed to know what kind of work to do next.

---

Boinc V 7.4.36
Win7 i5 3.33G 4GB NVidia 470

Still, I don't see why this is not a default action. The work has to get reported anyway.. and if the machine has completed an ENTIRE batch of work from one project, what harm is there in reporting it immediately?

I just can't see any harm it would do, and only see benefits. Not to mention, it -seems- like such an easy change considering how it handles this process now anyway.

---

Still, I don't see why this is not a default action. The work has to get reported anyway.. and if the machine has completed an ENTIRE batch of work from one project, what harm is there in reporting it immediately?

I just can't see any harm it would do, and only see benefits. Not to mention, it -seems- like such an easy change considering how it handles this process now anyway.

Because that the way the project developers want it to be.
I reiterate

This is so it can report batches of WU to the scheduler (which is a separate sever than the upload/download servers), and saves on bandwidth for you and the project.

---

Boinc V 7.4.36
Win7 i5 3.33G 4GB NVidia 470

Taken from the Wiki:

The reason for the two step policy is that the first stage only has to upload a file to a directory on the Data Server. The second stage requires a connection and contact with the BOINC Database. Because it is easier to report a list of files, the process was divided in this manner so that each part can be performed independently of the other.

If there is a heavy load on the server, if we tried to do both actions within the same transaction, any failure of the process invalidates the entire process. Since the file upload is a very simple transaction it can be easily performed independently of other activities.

---

I was about to query the same thing - I loaded SETI over ten days ago, and after an initial hiccup, it runs well. Not sure I like the "improved" graphics, so much!
The problem is that I ended up with four or five completed work units, and an indication that communication had been attemted to download the results which had failed. Rosetta and Einstein operated normally, ran, uploaded and downloade, but the completed SETI units sat there unnoticed and unwanted. Until today. They disappeared. So I checked the stats - no of work units completed - 0.00.
OOPS!
Secondary problem, probably unconnected, but machine now "freezes" when on standby and has to be reset using the reset button. It matters not whether machine set to go to standby after 10 minutes or 10 hours, it still freezes. This did not happen pre-BOINC. Problem is that it is then impossible to see what has happened as all messages are lost.
(AMD Duron 1.3, 224 Mb RAM, Win 98SE)

---

The messages are stored in stdoutdae.txt, in the BOINC folder.

Still, I don't see why this is not a default action. The work has to get reported anyway.. and if the machine has completed an ENTIRE batch of work from one project, what harm is there in reporting it immediately?

I just can't see any harm it would do, and only see benefits. Not to mention, it -seems- like such an easy change considering how it handles this process now anyway.

Because that the way the project developers want it to be.
I reiterate

This is so it can report batches of WU to the scheduler (which is a separate sever than the upload/download servers), and saves on bandwidth for you and the project.

You just answered the question "Why can't it report entirely complete batches of WU to the scheduler?"

with the answer... "This is so it can report batches of WU to the scheduler."

Excuse me if I either assume you 1. Didn't read what I'm talking about. 2. Are trying to confuse the hell out of even yourself. 3. are trying to upset me.

---

Taken from the Wiki:

The reason for the two step policy is that the first stage only has to upload a file to a directory on the Data Server. The second stage requires a connection and contact with the BOINC Database. Because it is easier to report a list of files, the process was divided in this manner so that each part can be performed independently of the other.

If there is a heavy load on the server, if we tried to do both actions within the same transaction, any failure of the process invalidates the entire process. Since the file upload is a very simple transaction it can be easily performed independently of other activities.

Here again we've got some space cadet telling me the reason for the two step process like he didn't read my post. I understand you think you're trying to help, but when you answer with stuff like that it's super fricking infuriating.

Guys.. I don't know how to make it any more clear, this is NOT WHAT I AM TALKING ABOUT.

I am completely aware of the two step process. Please go back and read my original post. Yep, It's down there at the bottom - WoW! Read it 3 or 4 times until it sinks in if you have to, I'm not talking about reporting each workunit when it's done. I'm talking about doing the SECOND stage of reporting, at the point in time after the entire batch of workunits have completed, instead of waiting additional time for god knows what, while BOINC queue's up work from AN ENTIRELY SEPERATE PROJECT. I'm not talking about tea time with the queen, I'm not talking about downloading work or changing my preferences, and I'm certainly not talking about reporting each individual workunit seperately. I understand this procedure in depth, and I most certainly don't need to be patronized or ingratiated by such rhetorical responses.

---

No, what you are talking about is the fact that you want to report immediately upon uploading. So all the answers we can give you for now is why it isn't done this way.

Uploading the file will write to a directory on the server.
Reporting will use the BOINC database of the project.

That may be fine if you are the only person uploading/reporting at the same time, but not if there's a couple of thousand of computers doing it at that same moment.
Using the database uses a lot more CPU & RAM on the server than writing a file to a directory does.

Hence why BOINC tries to spread this. So it's better that you report 2 or more files at the same time, not just 1.

The data is already in. The scientists who need this data can already look at what you sent them back.

The reporting is mostly done to get rid of the file in your download list and to give you a chance on credit.

Oh and please leave the "flaming" out of your posts. It's completely unnecessary.

---

No, what you are talking about is the fact that you want to report immediately upon uploading. So all the answers we can give you for now is why it isn't done this way.

NO IT ISN'T - READ IT AGAIN.

Then - READ IT AGAIN.

Here, I'll quote the important part.

"47 LHC (or SETI, or rosetta.. etc..) workunits I completed when they are all finished?"

Note, I mention 47 workunits. Not 1. - 47.

Then I say WHEN THEY ARE ALL finished.

If it had read - "47 LHC workunits when each one finishes" or "each LHC workunits when I completed them"

Then I could understand your seemingly oblivious remarks.

---

Since you don't want to read what the solution to your "problem" is, I don't want to read your posts any further. Officially ignored from now on.

And before you ask why it was done this way in Boinc 4.19, that's because it was a bug at that time. Have fun.

---

Since you don't want to read what the solution to your "problem" is, I don't want to read your posts any further. Officially ignored from now on.

And before you ask why it was done this way in Boinc 4.19, that's because it was a bug at that time. Have fun.

Good, cuz you're a freaking idiot. Thank god I don't have to put up with you screwing up my threads anymore!

---

Since you don't want to read what the solution to your "problem" is, I don't want to read your posts any further. Officially ignored from now on.

And before you ask why it was done this way in Boinc 4.19, that's because it was a bug at that time. Have fun.

Hey, I know this probably isn't helping, but I just had to comment on this one. (I normally tend to not comment on stuff like this, but can't help but see there is a communications shortfall here.) What you and Aaron here are talking about are two seperate things, man. I think you both know what you're talking about, but Aaron isn't asking what you think he is. He's talking about how BOINC can sometimes put off reporting even after the the total batch is done. Here, I'll try to illustrate for him.

Imagine for a second, a 2 SETI workunit cache. Very small, but that's all you got from the SETI project. Here's how it currently works sometimes.....

12:30 SETI Workunit 1(of 2 total) begins.
1:00 SETI Workunit 2(of 2 total) begins.
1:30 SETI Workunit 1 finishes, and uploads to the server.
1:30 LHC Workunit 1 begins.
2:00 SETI Workunit 2 finishes, and uploads to the server.
2:00 LHC Workunit 2 begins.
5:00 LHC Workunit 1 finishes, and uploads to the server.
5:00 LHC Workunit 3 begins.
5:45 The SETI Scheduler is contacted to REPORT the batch of WU's it completed at 2:00.

Now, As you can see, there is a 3 hour and 45 minute delay between the time when the batch is completed, and the client contacts the scheduler to report it.

What Aaron is asking is, Why can't we remove this 3 hour and 45 minute delay? I agree, and I can't see how you would read his statements any other way. By answering him with "Because reporting happens when the batch is completed" (as you have done repeatedly), is understandably frustrating and I hope you can see that - I personally would wanna smack ya. :-D I really find myself wondering if you did indeed, read his posts.

Anyway, I hope it's all explained now, and we can all go back to our little bowls of warm milk, curl up by the fire and mew ourselves to sleep like good kittens.

---

I think one problem is the term BATCH, who said that WUs are issued in BATCHES....?????

---

BOINC Wiki

Imagine for a second, a 2 SETI workunit cache. Very small, but that's all you got from the SETI project. Here's how it currently works sometimes.....

12:30 SETI Workunit 1(of 2 total) begins.
1:00 SETI Workunit 2(of 2 total) begins.
1:30 SETI Workunit 1 finishes, and uploads to the server.
1:30 LHC Workunit 1 begins.
2:00 SETI Workunit 2 finishes, and uploads to the server.
2:00 LHC Workunit 2 begins.
5:00 LHC Workunit 1 finishes, and uploads to the server.
5:00 LHC Workunit 3 begins.
5:45 The SETI Scheduler is contacted to REPORT the batch of WU's it completed at 2:00.

Now, As you can see, there is a 3 hour and 45 minute delay between the time when the batch is completed, and the client contacts the scheduler to report it.

Reporting is delayed so that there is a chance that multiple tasks can be reported at the same time. From your example above if there were no LHC tasks then there would be one more SETI task ready when the report occured.

---

BOINC WIKI

BOINCing since 2002/12/8

The problem seems to be in the order it reports work. Because it usually requests more work and reports while it is doing the uploads, the work it is uploading is not reported until the next time it requests more work. On an always-on connection this wouldn't be a big issue since you would only be a couple of work units behind. On a dial-up connection where you are only connected once or twice a day or less, the work would not be reported until the next connection - hours or days later.

I connect manually, let BOINC do all the uploading and downloading it needs to, and then manually update all projects with work to report before disconnecting. If BOINC would wait until all uploads were complete and ready to report before requesting new work, the extra step to reqort could be eliminated.

---

I completely agree with you Aaron - with your original post and with your subsequent rants about people not listening. [grin]

If you have "Network activity always available" then the BOINC manager should attempt to upload the result of the WU immediatly after completion (i.e., when it is "Ready to Report"). At the very least, there should be an option in the general preferences "Upload results immediatly ?".

The question then becomes: If the BOINC manager can't upload the results (for whatever reason), then how long should it wait before attemping again ?" Every minute ? every hour ? until the next WU is completed ? Again we could be given the option.

--- bt