Computation error
29 Nov 2016 10:30:55 UTC
Topic 203240
(moderation:
After a long while I opened Boinc manager and have found out that my Projects were stopped and I did not know when it had happened.
I updated them and after a while they started again but, while Pogs and Cas run the frocedure without problems, Einstein showed some:
1) it could not upload the results of the job it was working on before boinc stopped, so it could even be a few months. Since I knew that those results were useless for three other pc's had worked on the same parcel maybe last year, I pushed "Abort" and then "Update", after that the program downloaded some parcels of work but each one was stopped after 1 second and appeared "Computation error".
Has anyone an idea of where I can look for finding out a solution?
/var/log/boinc.log has, for Einstein,
29-Nov-2016 11:11:14 [Einstein@Home] update requested by user
29-Nov-2016 11:11:15 [Einstein@Home] Sending scheduler request: Requested by user.
29-Nov-2016 11:11:15 [Einstein@Home] Reporting 1 completed tasks
29-Nov-2016 11:11:15 [Einstein@Home] Requesting new tasks for CPU
29-Nov-2016 11:11:18 [Einstein@Home] Scheduler request completed: got 1 new tasks
29-Nov-2016 11:12:36 [Einstein@Home] Starting task h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0
29-Nov-2016 11:12:37 [Einstein@Home] Computation for task h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0 finished
29-Nov-2016 11:12:37 [Einstein@Home] Output file h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0_0 for task h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0 absent
29-Nov-2016 11:12:37 [Einstein@Home] Output file h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0_1 for task h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0 absent
29-Nov-2016 11:12:37 [Einstein@Home] Output file h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0_2 for task h1_0706.85_O1C02Cl1In0G__O1MD1G_G3473_707.00Hz_0_0 absent
/var/log/boincerr.log has a long list of
"No protocol specified"
Thanks,
Aldo
You are running Debian
)
You are running Debian Stretch with the 4.8 kernel that disables vsyscall and is currently preventing the GW app from running. We don't see problems with the FGRPB1 app so you might want to switch to that.
Please follow vsyscall is now disabled on latest Linux distros for updates on the situation.
Do you have a hostID?I
)
Do you have a hostID?
I tried to look at your computer but the account you used to post has no computers attached to it.
Edit: Christian was quicker, follow his advice. =)
Yes I have a HostID :
)
Yes I have a HostID : 188588
The problem is that last Spring I moved the hard disk to another computer and forget to check the situation of Boinc ! So yesterday, when I tried to have it starting again I erased from the list all the pc's I have used in the last 10 years and, obviously, since I hadn't sent yet any unity, the new one is not listed.
Christian, thank you for the link, I opened the Preferences and removed the tick from the following items:
Binary Radio Pulsar Search (Arecibo, GPU) - Boinc says that I do not have any usable GPU
Multi-Directed Continuous Gravitational Wave Search Tuning CV
Multi-Directed Continuous Gravitational Wave Search CVMulti-Directed Continuous Gravitational Wave Search G
Is that right?
Thanks :-)
Aldo
Turms wrote:Yes I have a
)
That is a very old hostID. It was probably created sometime in April of 2005. I have a hostID a little higher than that numerically that was created on 7th May 2005. Of course, mine is not the same hardware now. A 2005 vintage machine would be hard pressed to cope with today's tasks. I suspect your current machine is not that particular ID. A search of the database shows it's not currently allocated to any particular hardware or user. As Holmis points out, the user ID under which you are posting has no computers assigned to it. I suspect you have more than one account. You should log out of this account and login using the account that your current computer is associated with.
If you browse the BOINC data folder on your current host, you will find a file whose name is stdoutdae.txt which is the full BOINC event log for your current machine. It's a plain text file and can easily be browsed to find your current hostID (computer ID). The computer ID is listed as part of the startup messages whenever BOINC is restarted. If you have restarted BOINC recently, it might be easier to just scroll through the event log in BOINC manager.
Also in the BOINC data folder is another file whose name is account_einstein.phys.uwm.edu.xml which can also be browsed to show a 32 character <authenticator> which could be used to access the account for that host if you don't have the login details under which the host was registered. If you tell us the real computer ID from the startup messages, we should be able to tell you the account name it belongs to.
I guess that is why the hostID you gave is not now associated with any particular set of hardware. I don't delete any old hostIDs. I tend to reuse the ID for the machine that replaces it. Also, I sometimes like to jog my memory about long retired former hosts where the ID hasn't been reused.
Cheers,
Gary.
Thank you for your answer and
)
Thank you for your answer and time :-)
The <authenticator></authenticator> I have found in account_einstein..... ,
as for stdoutdae.txt it ends with:
22-May-2015 20:23:51 [---] Received signal 15
22-May-2015 20:23:51 [---] Exiting
obviously I do not remember if and why I launched a -TERM to boinc and, anyway, now boinc is running but I wonder why it is not uptodate.
and yes, as you say I started using boinc in 2005 mainly because I liked to help, for what I could, in the SETI project.
Just now, going through /var/log/boinc.log I have found:
30-Nov-2016 11:00:07 [Einstein@Home] URL http://einstein.phys.uwm.edu/; Computer ID 11674558; resource......
And also:
30-Nov-2016 15:33:06 [---] Account manager: BAM! User: 172325, Turms
Too many userid to cope with! :-D
... and I see that now my limit with Einstein is 2:
30-Nov-2016 13:27:55 [Einstein@Home] (reached daily quota of 2 tasks)
shouldn't it be 8? Maybe it is because, for some reason the Computation error is still there notwithstanding I've removed the mark from the tasks I listed in my previous message.
Thanks again,
Aldo :-)
Hi Aldo, I removed the
)
Hi Aldo,
I removed the authenticator from your post as it can be used to log into your account. As it was visible for some time you should consider this account to be compromised. Fortunately you still have your old account (which you use in the forums). The new account (which authenticator you posted) indeed has the host you are currently using and was created early October 2014 using a different email address.
I would advise you to remove Einstein@Home from this computer and reattach using the email address and password combination for your old account as this one is not compromised.
You can see the current email address of your account by clicking on "ACCOUNT" and then "Credentials". I can delete the compromised account once you attached your computer to your old account so you can even change your email address.
Christian, I thank you for
)
Christian, I thank you for your help, and I regret not having taken into account the risk in which I was incurring :-(
I'm going to act in the way you have suggested. Actually I do not think I'll reattach this computer for now since I'm of no help to the project just accepting tasks which are not performed. In a short while I move Debian to another disk because the present one has become noisy (and it is still Ata) and I fear that all of a sudden it will die!
Thank you again, and should I've problems, if you do not mind, I'll take the liberty of disturbing you.
Aldo :-)
I had some more problems
)
I had some more problems because boinc-manager doesn't start, I had to use boinccmd for detaching the project and reattaching it using the other mail address.
Now I have to study options of boinccmd to check if the computation error is still present or not.
BTW , boinc-manager doesn't start because of a mismatch between the c++ compiler used for the library and the one used for the program (at least this is what the error says!).