Hi, I love your idea, but I cannot add charityengine as account manager on BOINC v7.12.0 (x64) in Ubuntu 18.10. After manually entering your website and confirming credentials the event log states "Fetching configuration file from http://www.charityengine.com/get_project_config.php" and that's it! No error message, no progress and still the abilty to "Use account manager..." in Tools-Menu. What am I doing wrong?
To add to this question: I am trying to do a headless set-up of BOINC + Charity Engine on my Raspberry Pi and am also stuck with the account manager settings: the command
Similar question. I installed Boinc on a Ubuntu 18.04 box called 'skink'. That box had another identity ('localhost') some years ago, and another operating system.
The only project that's now appearing is "Fight Neglected Diseases", and no tasks are being found for that project.
So I suspect that an old record for this box is being accessed at the server end. How can we resolve this?
Ener and Nicolas, I don't see any particular reason for this to be happening, so we've turned on some extra logging for your accounts that may tell us more. If you try to attach to the account manager again, I can check to see if the logs shine any light on the problem.
Graham, same for you, I'll have to see what that box is sending to us to see why we aren't responding as expected. Is it possible it is connecting using a different account than yours?
Charity Engine | Sending scheduler request: Requested by user. Charity Engine | Requesting new tasks for CPU Charity Engine | Scheduler request completed: got 0 new tasks Charity Engine | No tasks sent Charity Engine | No tasks are available for ce9 Fight Neglected Diseases | Sending scheduler request: To fetch work. Fight Neglected Diseases | Requesting new tasks for CPU Fight Neglected Diseases | Scheduler request completed: got 0 new tasks
I'm not sure where the 'ce9' value comes from, but it's not my Account Name. Does that suggest anything?
No, ce9 is just the name of an application from the project server. I do think this confirms that this "skink" box is not attached to Charity Engine using your account, though. I turned on logging for your account and only see connections coming from another Linux box by a different name (and one that seems to be working fine).
Can you compare the account name (on the Projects tab in the advanced view) on these two systems to see if they match? If not, you could try detaching that box from Charity Engine and attaching again, using the same login that you use here on the Charity Engine website.
Tristan, it occured to me that CE may be mis-detecting the OS on skink because it's a pretty basic (Intel(R) Pentium(R) 4 CPU 3.00GHz) machine which can't do hardware virtualisation. So I shut down Boinc, removed virtualbox, and restarted Boinc. I then detached the Account Manager, removed all projects, and re-attached.
Tristan, did you have any more thoughts about the skink problem?
One other thing I noticed is that when CE is added as an Account Manager, it starts a Grid Republic Project which then starts and is replaced by the Fight Neglected Diseases Project.
Graham, I can see in your log file that the CE Desktop app is detecting Ubuntu 18.04.1 LTS and reporting that to the server, so no issue on your end there. It still isn't clear why that host isn't getting other projects, but I've made a change to see if it does anything (I'm thinking it's something in the way projects are automatically assigned, so we can try bypassing that).
Ener, would you be able to try installing a different version of BOINC? If not, there are other things I can look into, it will just take more time. I think 7.12.0 in particular included a number of account manager changes, so something may have broken there. If so, a newer or older version may work!
That would be my preference also, as that's the convenience of the package manager! What's strange is that I'm not seeing any connection from your cllient at all, as though it tries to connect, but never makes it to the server. Depending on your network environment, this could certainly be caused by a firewall on your system or between your system and the Internet.
Since it is just a HTTP request that would be blocked, and since you are able to access the Charity Engine website here without any problem, I'm only coming up with two possibilities, though:
a software firewall on Ubuntu is blocking outgoing HTTP / port 80 from BOINC, but allowing it from your web browser; or
a bug in BOINC is causing a problem when making the connection
That second possibility is what I'd suggest, since it's strange that the client doesn't give you an error (e.g. timeout), it just gets stuck.
Another thing to try first, though, without having to use a non-standard Ubuntu package, is to "Use account manager" again and enter https://www.charityengine.com instead of the usual http://www.charityengine.com. I have no reason to believe that HTTPS would work and HTTP wouldn't, but if there is a bug or a network issue only with HTTP, then it may be worth a shot.
@Tristan: I also thought about that and already tried to access charityengine via https. Unfortunately, I've got the same issue, so seems to be a bug in the Ubuntu version of BOINC.
But if it is really a bug, why isn't anyone else in the world faced with that?
Could be something like a localization issue, where it only fails when specific languages are used in the client, but the number of CE users on Ubuntu and BOINC 7.12.0 are not large in number, so it's hard to be certain! (Others are using this combination, but maybe they were already attached to CE and upgraded to 7.12.0 or the like, so they avoided the attachment bug that you're seeing?)
Actually the fix was in the command described from Nicolas in 8 Jan, Instead of using www.charitymanager.com I used www.charityengine.com
How can I be sure is actually doing something?
boinccmd --get_state shows I'm attached to charity engine project, but can't tell if it's downloading work or if it's just idle? Is there a command I can trigger to get jobs/work?
amazing idea but the implementation is really junk folks. Boinc is very stable and I've run Boinc since back in 2007.
After spending over a day trying to get the charityengine account manager to work with a linux install of Boinc I must say (out of frustration) were you folks dropped on your head when you implemented this ?
On windows it occasionally gets a task and runs MAX two cpu cores ever. sometimes its sitting there running a 0.0001 cpu core task and will not fetch any new work. I've installed this on six I7 PC's and one Ryzen PC
On Linux it does not matter how many times you reinstall and try join the account manager it does not fetch anything.
You're right, Linux support has not been good; and though we have taken steps to improve, it could still be better. Thanks for giving Charity Engine a try and sharing your experience with it.
For what it's worth, you will automatically be attached to more projects 3 days after joining (we do this so as not to burden research projects with accounts for users who don't stick around).
Hi, I love your idea, but I cannot add charityengine as account manager on BOINC v7.12.0 (x64) in Ubuntu 18.10. After manually entering your website and confirming credentials the event log states "Fetching configuration file from http://www.charityengine.com/get_project_config.php" and that's it! No error message, no progress and still the abilty to "Use account manager..." in Tools-Menu. What am I doing wrong?
To add to this question: I am trying to do a headless set-up of BOINC + Charity Engine on my Raspberry Pi and am also stuck with the account manager settings: the command
boinccmd --join_acct_mgr www.charitymanager.com *e-mail* *password*
just returns
poll status: operation in progress
and then nothing else. Seems like a similar problem to the one described above.
Similar question. I installed Boinc on a Ubuntu 18.04 box called 'skink'. That box had another identity ('localhost') some years ago, and another operating system.
The only project that's now appearing is "Fight Neglected Diseases", and no tasks are being found for that project.
So I suspect that an old record for this box is being accessed at the server end. How can we resolve this?
Ener and Nicolas, I don't see any particular reason for this to be happening, so we've turned on some extra logging for your accounts that may tell us more. If you try to attach to the account manager again, I can check to see if the logs shine any light on the problem.
Graham, same for you, I'll have to see what that box is sending to us to see why we aren't responding as expected. Is it possible it is connecting using a different account than yours?
Here's what I'm seeing in the logs:
Charity Engine | Sending scheduler request: Requested by user.
Charity Engine | Requesting new tasks for CPU
Charity Engine | Scheduler request completed: got 0 new tasks
Charity Engine | No tasks sent
Charity Engine | No tasks are available for ce9
Fight Neglected Diseases | Sending scheduler request: To fetch work.
Fight Neglected Diseases | Requesting new tasks for CPU
Fight Neglected Diseases | Scheduler request completed: got 0 new tasks
I'm not sure where the 'ce9' value comes from, but it's not my Account Name. Does that suggest anything?
No, ce9 is just the name of an application from the project server. I do think this confirms that this "skink" box is not attached to Charity Engine using your account, though. I turned on logging for your account and only see connections coming from another Linux box by a different name (and one that seems to be working fine).
Can you compare the account name (on the Projects tab in the advanced view) on these two systems to see if they match? If not, you could try detaching that box from Charity Engine and attaching again, using the same login that you use here on the Charity Engine website.
OK, I've detached and re-attached. The log file an image of the Charity Image screenshot are at: https://we.tl/t-Ovc3x4vnYO
The latter indicates that CE may not be detecting that skink is a Linux Ubuntu box. But it is:
lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 18.04.1 LTS
Release: 18.04
uname -a
Linux skink 4.15.0-43-generic #46-Ubuntu SMP Thu Dec 6 14:45:28 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
So how do we fix this?
Tristan: Just tried it again. Same as described in my first post: Fechting config... nothing else happens...
Tristan, it occured to me that CE may be mis-detecting the OS on skink because it's a pretty basic (Intel(R) Pentium(R) 4 CPU 3.00GHz) machine which can't do hardware virtualisation. So I shut down Boinc, removed virtualbox, and restarted Boinc. I then detached the Account Manager, removed all projects, and re-attached.
Same result :(
Tristan, did you have any more thoughts about the skink problem?
One other thing I noticed is that when CE is added as an Account Manager, it starts a Grid Republic Project which then starts and is replaced by the Fight Neglected Diseases Project.
So where do we go from here??
Graham, I can see in your log file that the CE Desktop app is detecting Ubuntu 18.04.1 LTS and reporting that to the server, so no issue on your end there. It still isn't clear why that host isn't getting other projects, but I've made a change to see if it does anything (I'm thinking it's something in the way projects are automatically assigned, so we can try bypassing that).
Ener, would you be able to try installing a different version of BOINC? If not, there are other things I can look into, it will just take more time. I think 7.12.0 in particular included a number of account manager changes, so something may have broken there. If so, a newer or older version may work!
Yes, that worked .. see screenshots at: https://we.tl/t-cZRJAllpWH
But Fight Neglected Diseases is still in the Project list. And Rosetta isn't.
Anyway, I've now got something that works, thanks. It's only a test machine, but I'm hoping to upgrade it sometime soon. Thanks for your help!
@Tristan: I try to run as far as possible with the standard Ubuntu packages, so - if you agree - I'd like to stay on the current version.
That would be my preference also, as that's the convenience of the package manager! What's strange is that I'm not seeing any connection from your cllient at all, as though it tries to connect, but never makes it to the server. Depending on your network environment, this could certainly be caused by a firewall on your system or between your system and the Internet.
Since it is just a HTTP request that would be blocked, and since you are able to access the Charity Engine website here without any problem, I'm only coming up with two possibilities, though:
That second possibility is what I'd suggest, since it's strange that the client doesn't give you an error (e.g. timeout), it just gets stuck.
Another thing to try first, though, without having to use a non-standard Ubuntu package, is to "Use account manager" again and enter https://www.charityengine.com instead of the usual http://www.charityengine.com. I have no reason to believe that HTTPS would work and HTTP wouldn't, but if there is a bug or a network issue only with HTTP, then it may be worth a shot.
@Tristan: I also thought about that and already tried to access charityengine via https. Unfortunately, I've got the same issue, so seems to be a bug in the Ubuntu version of BOINC.
But if it is really a bug, why isn't anyone else in the world faced with that?
Could be something like a localization issue, where it only fails when specific languages are used in the client, but the number of CE users on Ubuntu and BOINC 7.12.0 are not large in number, so it's hard to be certain! (Others are using this combination, but maybe they were already attached to CE and upgraded to 7.12.0 or the like, so they avoided the attachment bug that you're seeing?)
Hello, im facing the same problem, infinite loop of
polling for reply
poll status: operation in progress
is there a specific version where the account manager works in ubuntu? im currently running on ubuntu server
lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 19.04
Release: 19.04
Codename: disco
boinc client is version 7.14.2+dfsg-3
thanks in advance
You could try a development release to see if it is any better, though this would require adding a PPA via apt:
sudo add-apt-repository ppa:costamagnagianfranco/boinc
sudo apt update
sudo apt install boinc
https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/boinc
Unfortunately I'm unable to try this myself, as my version of Ubuntu is too old!
Actually the fix was in the command described from Nicolas in 8 Jan, Instead of using www.charitymanager.com I used www.charityengine.com
How can I be sure is actually doing something?
boinccmd --get_state shows I'm attached to charity engine project, but can't tell if it's downloading work or if it's just idle? Is there a command I can trigger to get jobs/work?
My setup is command line only
Sure, there are plenty of boinccmd options; in this case:
amazing idea but the implementation is really junk folks. Boinc is very stable and I've run Boinc since back in 2007.
After spending over a day trying to get the charityengine account manager to work with a linux install of Boinc I must say (out of frustration) were you folks dropped on your head when you implemented this ?
On windows it occasionally gets a task and runs MAX two cpu cores ever. sometimes its sitting there running a 0.0001 cpu core task and will not fetch any new work. I've installed this on six I7 PC's and one Ryzen PC
On Linux it does not matter how many times you reinstall and try join the account manager it does not fetch anything.
You could not have screwed this up any worse.
p.s. This is my profile on WorldCommunityGrid: https://www.worldcommunitygrid.org/stat/viewMemberInfo.do?userName=Tikiri%20Wickramasingha
9/11/2019 1:41:37 am | Charity Engine | update requested by user
9/11/2019 1:41:40 am | Charity Engine | Sending scheduler request: Requested by user.
9/11/2019 1:41:40 am | Charity Engine | Requesting new tasks for CPU
9/11/2019 1:41:41 am | Charity Engine | Scheduler request completed: got 0 new tasks
9/11/2019 1:41:41 am | Charity Engine | No tasks sent
9/11/2019 1:41:51 am | Charity Engine | Sending scheduler request: To fetch work.
9/11/2019 1:41:51 am | Charity Engine | Requesting new tasks for CPU
9/11/2019 1:41:52 am | Charity Engine | Scheduler request completed: got 0 new tasks
9/11/2019 1:41:52 am | Charity Engine | No tasks sent
You're right, Linux support has not been good; and though we have taken steps to improve, it could still be better. Thanks for giving Charity Engine a try and sharing your experience with it.
For what it's worth, you will automatically be attached to more projects 3 days after joining (we do this so as not to burden research projects with accounts for users who don't stick around).
Ohhhhhhhh why didn't you say that :)
So I should just leave it running for a few days ?
Or should I start this 3 days after joining.
Its a total of:
5 X I7 4790 = 8 threads @3.4GHz X 5
1 X I7 4790K = 8 threads @ 4.4 GHz
1 X Ryzen 2700 = 16 threads @ 3.4 Ghz
16 Cores on Azure = 16 E5 threads @ 3.2 GHz
Plus currently powered off: Three servers 128GB ram each with a total of 64 + 48 + 24 Cores = 136 cores.
Linux systems are not likely to receive work until it's been three days. It would be interesting to hear how it goes at that point!