Across-projects Team

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 94,820,839
RAC: 122,607

Thanks for explanations! If

Thanks for explanations! If understand right, actually, in BOINC-projects user have one "cross-projects key", that identify it for global preferences and statistics - user e-mail. :(

> I had to use a different email account to set up CPDN (and it had a different
> CPID from the other projects), and what was happening was that my CPDN general
> preferences were out of sync with the other projects, so every time my
> machines would connect to CPDN, the CC would change to those, and when they
> would connect to other projects, they would revert back. I had to go log in to
> CPDN separately and make sure all of the general preferences were the same as
> at the other projects.

> Once they are all the same address, then your CPID will be the same for each
> project (it may take a couple of days to propagate) and your preferences will
> stay in sync.

I have a same situation. I create an BOINC-CPDN account with using a different e-mail, (different from Einstein@Home). (I didn't know about "equivalent e-mail" restriction). And third day I don't receive the letter for confirm of e-mail changing! :(

Why BOINC developers don't create single database on all BOINC-users and teams?????? WHY???

One database, one primary key = one record = one user = one preferences = one solid statistic (total statistics and statistics on different projects), only one, not project, but BOINC-accont with only "attach to project" action to starting in any project!

Why my e-mail used as a primary key? (actually used).

Decentralization, of course, is a good thing, but in this case it is not good for users and statistics. And centralization is a strongly required...

Keck_Komputers
Keck_Komputers
Joined: 18 Jan 05
Posts: 376
Credit: 5,744,955
RAC: 0

>Why BOINC developers don't

>Why BOINC developers don't create single database on all BOINC-users and
>teams?????? WHY???

BOINC is a distributed computing framework. It has no business keeping any type of central database. They are working on a means of attaching to multiple projects in a single step. Even that may be too much centralization.

Centralization could limit the ability of projects to use specialized clients like those japaneese projects do. It could also make the system more suseptable to one project going rouge and making a mess of all projects.

BOINC WIKI

BOINCing since 2002/12/8

[AF>Linux]Arnaud
[AF>Linux]Arnaud
Joined: 22 Jan 05
Posts: 19
Credit: 27,188
RAC: 0

Just a stupid question: I

Just a stupid question:
I have 2 CPID and no prob of stats: Do you know why and is there a mean to merge them ? (I always use the same e-mail address, of course)
1 CPID for seti,E@H,LHC,P@H,Burp and the other one for CPDN, F@H, and other Alpha projects.

Arnaud

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 94,820,839
RAC: 122,607

>(edit) IF you have set up

>(edit) IF you have set up different email addresses for different projects, it
> is a simple matter to go and edit them in your accounts. You will then have to
> receive and respond to a confirmation email from each project you change it
> in.
>
> Once they are all the same address, then your CPID will be the same for each
> project (it may take a couple of days to propagate) and your preferences will
> stay in sync.

It is true or false? (I don't know because I started in CPDN, but have a problem with edition of account e-mail :( )

If it is true, single BOINC-users database already exist. Database for user preferences. Consequently, if other projects can synchronise global preferences, its have an access to this database. Any database must have a field for unique key. Consequently, questions about forms of attaching to projects depends only from organisation of projects.

But, of course it is only my opinion. Reality may be more interesting. :)

MJKelleher
MJKelleher
Joined: 20 Feb 05
Posts: 7
Credit: 36,534,580
RAC: 19,545

I believe multiple

I believe multiple cross-project IDs can happen when there are multiple computers which crunch different projects. If one computer only runs CPDN and SETI, and another runs Einstein and Protein, there is no opportunity for information to get in synch among the projects and computers. When I run SETI on all my (three) computers, my project information, and (I believe) the CPID get synched up, with the most recent information replacing the earlier. It can take a little time, multiple updates, for the information to reach all computers and all projects.

BOINC Wiki Thanks to Paul D. Buck and associates

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

> If it is true, single

Message 10093 in response to message 10091

> If it is true, single BOINC-users database already exist. Database for user
> preferences. Consequently, if other projects can synchronise global
> preferences, its have an access to this database. Any database must have a
> field for unique key. Consequently, questions about forms of attaching to
> projects depends only from organisation of projects.

There is no cross-project database. The CPID, your accounts and your clients communicating with the Schedulers is how the projects get synchronized. It is also why it does not happen that fast (usually).

Especially when you have a long time in your buffer settings ...

Vid Vidmar*
Vid Vidmar*
Joined: 22 Jan 05
Posts: 25
Credit: 191,816
RAC: 0

> Just a stupid question: > I

Message 10094 in response to message 10090

> Just a stupid question:
> I have 2 CPID and no prob of stats: Do you know why and is there a mean to
> merge them ? (I always use the same e-mail address, of course)
> 1 CPID for seti,E@H,LHC,P@H,Burp and the other one for CPDN, F@H, and other
> Alpha projects.

Hi. It's not stupid question at all. Fact is, as mentioned before in this thread, that there are 2 CPIDs. CPDN and majority of alpha projects show the "internal" CPID on account page(s), but for XML export use the "external" CPID which is shown on other project's account page(s). Example: my CPID is: 85ca0bcd25d470318dcaecbe8029788c, CPDN (and some other projects) show: c81cda9ddc306618a617f763b6d9d95b.

Hope, this explains your question.
Happy crunching,

[AF>Linux]Arnaud
[AF>Linux]Arnaud
Joined: 22 Jan 05
Posts: 19
Credit: 27,188
RAC: 0

> Hope, this explains your

Message 10095 in response to message 10094

> Hope, this explains your question.
> Happy crunching,
>
Yes, it explains perfectly :o)
Thanks and happy crunching

Arnaud

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 94,820,839
RAC: 122,607

> There is no cross-project

Message 10096 in response to message 10093

> There is no cross-project database. The CPID, your accounts and your clients
> communicating with the Schedulers is how the projects get synchronized. It is
> also why it does not happen that fast (usually).
>
> Especially when you have a long time in your buffer settings ...

It's make situation more clear... Thank you!

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 94,820,839
RAC: 122,607

Example: One user have a two

Example: One user have a two computers - one for CPDN and another - for Einstein@Home.
Both accounts created with using same e-mail address, but absolutely separately. First computer only for CPDN, and another computer - only for Einstein. What "say" the statistic systems about this situation?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.