OK... but Trac doesn\'t use the same anti-spam code. Only the forums do. Funny detail, these CPDN forums were using the code before the BOINC forums did.

Ah, never mind. I see the ticket. Will assign it to the right person.

---

Jord.

OK... but Trac doesn\'t use the same anti-spam code.

Trac uses a different anti-spam code, one comparing across a couple of servers, including Akismet.

Anyway, I just learned from Rom that this will be fixed in a couple of days, when the code has learned from its mistakes. Trac needs some 15 more tickets with logs to learn they aren\'t hostile.. In the mean time I assigned your ticket to Rom. he asked for it.

---

Jord.

Les wrote:

The real problem is if a person has the problem and let\'s it go for days before saying something here, so that we can tell Milo.
After 14 days of attempting to upload a zip file, BOINC will give up and delete it.
This has already happened to at least one person.

Oh no!

Best regards, MM

MikeMarsUK wrote:

Thyme Lawn suggested some ideas about how to get a failed upload to be resent, so I\'m going to experiment and see if I can write a programme to do this. Obviously it wouldn\'t work if Boinc has already deleted it!

Thanks Mike - but at the moment, both my BBC models try (unsuccessfully) to upload again every time I resume network activity.

From the stdoutdae file, I can see that the problem on this comp started with a trickle in the early hours of Wed, when it completed 2060. My other BBC model had its first upload problem that same afternoon when it trickled 2010.

Three (that I know of) of my team-mates are having the same problem in CPDN - one trying to upload 1930, one trying to report a completed model, and one a crashed model. However, my own CPDN model has just trickled 1968 without incident. So it rather looks as if, in either BBC or CPDN the problem is only with the 10-yearly data dumps or with the final report (whether due to completion or crash).

(BTW someone has been \"minusing\" posts in this thread again!)

---

Visit the Scotland team

Hi MM

The \'fix\' talked about by Mike And Thyme Lawn, is for AFTER the 14 day period, if it comes to that.
It CAN\'T work before the servers are OK again, because the servers aren\'t accepting data.
And if they start accepting data before then, the fix won\'t be needed.

It all depends on where in the 10 year period people are.
E.G. Mine is in 2033, so I\'m hoping that it will be fixed by then. And if not, then I\'ll suspend it before it gets that far and just wait.

There are several upload servers, and it can be \'the luck of the draw\' whether you get one with enough space for your upload. But having said that, they\'re all crammed to the gills with data (terabytes of the stuff).

Regarding the minuses which are turning up on the forum, I\'d advise everyone to simply ignore them - they\'re harmless and a bit sad. If someone has valid feedback (positive or negative), the constructive thing to do is to express it is to actually post (or PM), rather than clicking on every \'-\' link in sight.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

The slightly annoying thing about the full servers is that every time I turn networking back on for the benefit of other projects, my machine uploads 10MB (soon to be 15MB) first, and only gets the message that there\'s no space afterwards. Not good for bandwidth - yours or mine.

I\'m sure that BOINC signals the size of the file it\'s uploading before the transfer even starts: could someone put in a feature request for the next version of BOINC (I\'m at v5.8.16) that would allow it to check for server space before the transfer, and send back a backoff message if there\'s a problem?

There\'s a link for submitting requests for Boinc development:

http://boinc.berkeley.edu/trac/newticket

(The new Boinc/TRAC system).

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

I know what needs to be changed in the client and file upload handler, so I\'ll create a ticket.

Edit: created here

---

"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

Cheers, thanks for that :-)

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

That sounds like a really good idea especially for multi-project crunchers. The ticket is #133.

---

Cpdn news

It will do no harm at all just to suspend boinc network activity and let these 10-year zip files wait a day or two to upload.

i have currently suspended bonic network activity on two computers, the main issue i am having with doing this is i do like to be able to crunch other projects as well, basically i run cpdn work units as a back up to other projects in case they go down which seems to have back fired some what since i am now only crunching back up project rather then then main ones i want to be crunching and i am basically being forced into this position because i have bandwidth limit on internet access

also some sort of idea of time frame it will take to clear up this issue would be nice since i have been having this issue for the last couple of days and was suggested it would be resolved in a day or two

Does suspending CPDN help? (I can\'t remember if it suspends the uploads or not, I\'m not near a PC with Boinc).

Resolving the problem involves ordering new disks and installing them into the servers, I know the disks were ordered a few days ago but don\'t know when they\'re arriving / being fitted. Hopefully very soon.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Does suspending CPDN help? (I can\'t remember if it suspends the uploads or not, I\'m not near a PC with Boinc).

Suspending BOINC won\'t suspend ongoing uploads.
Suspending Network activity might... Although the latest versions (5.8.x) will try for another 5 minutes after putting this option on to have the uploads continue.

---

Jord.

I know what needs to be changed in the client and file upload handler, so I\'ll create a ticket.

Edit: created here

Thanks for doing that (and for the feedback, Mike and Mo).

I always like to put these random suggestions up for open discussion, rather than going straight to trac or equivalent - if it gets general approval, it feels less likely that I\'m just expecting BOINC to be moulded to my personal whims. Also, Thyme put together a much better-worded ticket than I could ever have done.

Does suspending CPDN help? (I can\'t remember if it suspends the uploads or not, I\'m not near a PC with Boinc).

Resolving the problem involves ordering new disks and installing them into the servers, I know the disks were ordered a few days ago but don\'t know when they\'re arriving / being fitted. Hopefully very soon.

i have tried that and boinc still tries to upload results (currently using 5.4.11) with boinc view 1.2.6 i have the option of aborting transfer which i guess i shouldnt be using ie it would only bring the 14 day period to now , should i copy zip files to another location just in case and if so how would i go about getting zipped files to cpdn ?

thank you for update on resolution to issue

i have the option of aborting transfer which i guess i shouldnt be using ie it would only bring the 14 day period to now

No, what it does is abort the whole transfer. So whatever you do, don\'t use this option.

---

Jord.

Hi Dynamite

As far as I know, when you allow network activity, this always allows ALL projects with trickles or files for their servers to try to send them.

We\'re advising members to suspend network activity to prevent indefinite quantities of worrying boinc manager messages when really at the moment most members have no problem. Waiting for a few days to contact the server - now probably until at least Monday or Tuesday - is not a problem.

But if you crunch more than one project, you can still allow network activity for long enough to send your trickles and upload files and get new work. You should then suspend network activity again. (You will see from the boinc manager messages that your other project(s) have finished talking to the server.) This isn\'t an ideal solution - we\'re asking you to control network access manually through the Activity menu - but it does allow you to report workunits and get new work from other projects.

As an alternative solution, you could set your other projects to No new work at the moment and let cpdn run all the time with network activity suspended while this problem lasts. When the server problem is resolved, you could suspend your cpdn model for an equal number of days to restore your balance.

---

Cpdn news

One more suggestion which may help a little bit with your other projects:

Make a note of how long you get on deadline for each project you attach to. For example, I think Seti gives 10 days.

Take one third of this (3 days for the 10 day example above), and then go to :

* Your account
* General preferences
* Network usage / Connect to network about every:

Put the 3 days into the connection frequency box, save, and update your client.

This will mean that every time you attach, you get 3 days work from each project, and you don\'t need to reattach for that 3 days.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Put the 3 days into the connection frequency box, save, and update your client.

This will mean that every time you attach, you get 3 days work from each project, and you don\'t need to reattach for that 3 days.

Again being a devil in a shoebox... that will only work for older BOINC clients.
BOINC 5.8.x will only request 1 second of work when you attach. Then it will read the preferences and only then will it ask for more work, spread over an amount of time. To get that amount of work in, it can take about 30 minutes, if not more. Depends on the activity at the server at that time.

---

Jord.

If people already have the failing upload problem, then it\'s too late to do much about it.

For those APPROACHING a 10 year point in their crunching, I\'d recommend that:
1) Make sure that the project is set to \"No new tasks\" in the Projects tab
2) Suspend the model in the Tasks tab
3) Wait until the problem is resolved
4) Then restart the model.

---

Backups: Here