I'm having this issue again on at least one of my hosts--will check others, but of the two I've checked only the OS X 10.8 host has suffered from this, but not Windows 7.

The issue is that, after changing my work preferences recently via BAM, a "wrong" set of work preferences is sent by the projects even though they are attributed to BAM. The major error in these "wrong" preferences, is that the "don't compute while active" flag is set in the wrong preferences when it is not set to do so in BAM. Syncing directly with BAM sends the correct work preferences, but they are overridden as soon as a project is updated. I also upgraded the hose in question from BOINC 7.2.33 to 7.2.42 and that made no difference. See log example below:

Sun Apr 20 18:05:47 2014 | | Contacting account manager at http://bam.boincstats.com/
Sun Apr 20 18:05:49 2014 | | Account manager: BAM! User: 13859, noderaser
Sun Apr 20 18:05:49 2014 | | Account manager: BAM! Host: 419603
Sun Apr 20 18:05:49 2014 | | Account manager: Number of BAM! connections for this host: 512
Sun Apr 20 18:05:49 2014 | | Account manager: Dear founder of team HSI Production Services, your team is invited to participate in the team challenges (http://boincstats.com/en/bam/teamChallenge/)
Sun Apr 20 18:05:49 2014 | | Account manager contact succeeded
Sun Apr 20 18:05:49 2014 | | General prefs: from http://bam.boincstats.com/ (last modified 20-Apr-2014 00:55:33)
Sun Apr 20 18:05:49 2014 | | Host location: none
Sun Apr 20 18:05:49 2014 | | General prefs: using your defaults
Sun Apr 20 18:05:49 2014 | | Preferences:
Sun Apr 20 18:05:49 2014 | | max memory usage when active: 9830.40MB
Sun Apr 20 18:05:49 2014 | | max memory usage when idle: 16384.00MB
Sun Apr 20 18:05:49 2014 | | max disk usage: 50.00GB
Sun Apr 20 18:05:49 2014 | | don't use GPU while active
Sun Apr 20 18:05:49 2014 | | (to change preferences, visit a project web site or select Preferences in the Manager)


Sun Apr 20 18:28:08 2014 | WUProp@Home | Sending scheduler request: Requested by project.
Sun Apr 20 18:28:08 2014 | WUProp@Home | Not requesting tasks: non CPU intensive
Sun Apr 20 18:28:10 2014 | WUProp@Home | Scheduler request completed
Sun Apr 20 18:28:10 2014 | | General prefs: from http://bam.boincstats.com/ (last modified 20-Apr-2014 00:56:19)
Sun Apr 20 18:28:10 2014 | | Host location: none
Sun Apr 20 18:28:10 2014 | | General prefs: using your defaults
Sun Apr 20 18:28:10 2014 | | Preferences:
Sun Apr 20 18:28:10 2014 | | max memory usage when active: 8192.00MB
Sun Apr 20 18:28:10 2014 | | max memory usage when idle: 14745.60MB
Sun Apr 20 18:28:10 2014 | | max disk usage: 197.87GB
Sun Apr 20 18:28:10 2014 | | don't compute while active
Sun Apr 20 18:28:10 2014 | | don't use GPU while active
Sun Apr 20 18:28:10 2014 | | suspend work if non-BOINC CPU load exceeds 25%
Sun Apr 20 18:28:10 2014 | | (to change preferences, visit a project web site or select Preferences in the Manager)
Sun Apr 20 18:28:11 2014 | | Suspending computation - computer is in use
Sun Apr 20 18:28:11 2014 | | Suspending network activity - computer is in use
Sun Apr 20 18:28:28 2014 | | Suspending GPU computation - computer is in use


WUProp is just one example project, but the "bad" preferences seem to come from any project. Also, the projects report that the preferences have come from BAM--is there something wrong with the way BAM is synching its work preferences? Why would it only affect (so far) this one host?

This is not a client issue. The preferences are synced directly between BAM! and the projects. Since the project indicates the correct modify time for the preferences we can assume that the sync still works. There does seem to be a problem with getting the correct preferences to the project. At BAM!'s side the code has not changed since August last year. I'll have to ask the BOINC devs if the webcode has changed.

It's odd that I don't encounter this issue on all of my hosts, only some of them. I have seen such behavior on a single host for a short period of time (I'm guessing preference sync errors that are fixed by subsequent updates) but this is the first time I have seen it on multiple hosts simultaneously.

I notice that in both the BAM and project messages, it says my host location is "none", while I have locations set for all of my hosts on the BAM Host list.

I find it odd that I have one host that is not experiencing the issue, a Win7 64-bit host running 7.2.33, even though another host with same OS and BOINC version is affected. Here's a log from the "immune" host:

4/20/2014 11:52:34 PM | | Fetching configuration file from http://bam.boincstats.com/get_project_config.php
4/20/2014 11:52:38 PM | | Contacting account manager at http://bam.boincstats.com/
4/20/2014 11:52:40 PM | | Account manager: BAM! User: 13859, noderaser
4/20/2014 11:52:40 PM | | Account manager: BAM! Host: 265184
4/20/2014 11:52:40 PM | | Account manager: Number of BAM! connections for this host: 893
4/20/2014 11:52:40 PM | | Account manager: Dear founder of team HSI Production Services, your team is invited to participate in the team challenges (http://boincstats.com/en/bam/teamChallenge/)
4/20/2014 11:52:40 PM | | Account manager contact succeeded
4/20/2014 11:52:40 PM | | General prefs: from http://bam.boincstats.com/ (last modified 20-Apr-2014 00:55:33)
4/20/2014 11:52:40 PM | | Host location: none
4/20/2014 11:52:40 PM | | General prefs: using your defaults
4/20/2014 11:52:40 PM | | Reading preferences override file
4/20/2014 11:52:40 PM | | Preferences:
4/20/2014 11:52:40 PM | | max memory usage when active: 4095.24MB
4/20/2014 11:52:40 PM | | max memory usage when idle: 7371.43MB
4/20/2014 11:52:40 PM | | max disk usage: 279.34GB
4/20/2014 11:52:40 PM | | don't use GPU while active
4/20/2014 11:52:40 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)
4/20/2014 11:52:49 PM | WUProp@Home | update requested by user
4/20/2014 11:52:52 PM | WUProp@Home | Sending scheduler request: Requested by user.
4/20/2014 11:52:52 PM | WUProp@Home | Not requesting tasks: non CPU intensive
4/20/2014 11:52:59 PM | WUProp@Home | Scheduler request completed
4/20/2014 11:52:59 PM | | General prefs: from http://bam.boincstats.com/ (last modified 20-Apr-2014 00:56:19)
4/20/2014 11:52:59 PM | | Host location: none
4/20/2014 11:52:59 PM | | General prefs: using your defaults
4/20/2014 11:52:59 PM | | Reading preferences override file
4/20/2014 11:52:59 PM | | Preferences:
4/20/2014 11:52:59 PM | | max memory usage when active: 4095.24MB
4/20/2014 11:52:59 PM | | max memory usage when idle: 7371.43MB
4/20/2014 11:52:59 PM | | max disk usage: 279.34GB
4/20/2014 11:52:59 PM | | don't use GPU while active
4/20/2014 11:52:59 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)

Any updates?

I still seem to be having some issues with preferences and host location; things seemed to be fine for a while, but I am having the issue with the mysterious "suspend work if non-BOINC CPU load exceeds 25%" again. All hosts also report as being "Host location: none" when syncing with BAM, even though a host location is set at both BAM and the projects.

Since nobody else is reporting this problem, could there be something corrupted with my account?

Hi,

I think I have the same issue on my Ubuntu Linux client, reported as well in the Boinc forums: Link

Today I managed to get a new and clean installation of my client to demonstrate the problem. After a complete removal of the client with

and manual removal of "/etc/boinc-client", then machine reboot, I performed an clean installation of the client:


Then I attached Asteroids@Home with this command:


and checked the success with


--> everything looked successful. So I decided to have a look into "stdoutdae.txt":


First question: Why is the client connecting to Bam and taking settings from there? I did not attach the client to Bam with the command " --join_acct_mgr URL name passwd"! I expected the software to use the prefs from Asteroids@Home...

Second question: I have set up some work preferences in Bam previously for this client which allow the client to compute while active. Why is the client not using these prefs when connecting to Bam?

I've never had such problems with my Windows 7 clients or with those running under Android...

Any ideas?

Best wishes,
Alex

I had very similar problem with my CentOS hosts. I installed boinc clients, connected them to BAM!, then created custom Work Perferences with specific variables like "not computing if CPU usage is higher as 60%" and attached hosts to this.

Unfortunately strange things appear. After synchronisation with BAM! hosts not always worked correctly. Sometimes it suspended network activity, sometimes CPU activity (not because of CPU usage). When I tried to resync hosts with BAM! it helped for short time only.

Because I am not very familar with this OS, my choice was to use local global_prefs_override.xml files with perferences I need. After this step, hosts working without any problems. So, maybe it was my problem with lack of knowledge, but also it is possible, that in BAM! are some bugs.

I never tried BAM! Work Perferences on Windows OS.

have the same error seems like thats OK?

OrangePI running Ubuntu 14.04.3 with BOINC 7.2.42

Edit: seems like I resolved it ... one Project had an preference set to "home" I had to unset it via the Project webinterface stop the daemon remove all account_*, master_*, sched_* and statistic_* files
Restart the Client, it synces the files again and than the Pref isnt set and it uses the correct BAM setting !

Because the projects and BAM! share the same preferences "pool". The project (or BAM!) that has the latest updated preferences will have it's preferences pushed to all the other projects through the client. Also, BAM! directly writes the preferences to all projects when they are changed. Therefor you don't have to attach to BAM! to get the BAM! prefs if BAM! has the newest prefs.


These prefs can only be send to the host when it is attached to BAM!.