Problem with the upload of processed data.
18 Nov 2012 4:43:56 UTC
Topic 196633
(moderation:
Data of about 4-5 projects have been processed on my system up till now but the problem is the final processed data doesn't get uploaded.
I have installed BOINC manager on my system.
Any help in this respect will be appreciated.
You can reply back to me @ (if it's not a problem for you).
Thanks.
Problem with the upload of processed data.
)
This project had an outage yesterday. See this thread for details on that.
After they came back, they kept all systems down, to thoroughly check them. So that's why you couldn't contact them. However, now the project is back up, you can upload, you can report.
If BOINC is still in a form of project back-off, just wait it out. It'll upload & report eventually (and within the next 24 hours).
Edit: Although, after checking your contact data, it seems your BOINC hasn't tried to make contact with the project since the 4th of November. So either something is blocking BOINC from making contact with the project servers, think about a firewall or some sort. Or you set BOINC to not use the network any further. That's something you can check in BOINC Manager.
Sir, I tried doing away the
)
Sir, I tried doing away the firewalls but it didn't help.
Any other suggestions??
RE: Sir, I tried doing away
)
Your thread title suggests you have a problem with uploading completed results but that cannot be true because you don't have any completed results to upload??
You said that you had processed about 4-5 "projects" but I think you meant "tasks" since you seem to be attached only to the E@H project on your computer. You received just two tasks on 4th November and you aborted them on 18th November. For the server to know this, your host must have successfully uploaded information to the server so there isn't (at least now) a problem with uploading.
Why do you think your computer had finished processing the data for those two tasks? The aborted results that were sent back show that one task had accumulated only about 65 seconds of CPU time and the other had accumulated zero. You couldn't possibly have seen 100% completed tasks in your BOINC Manager in such a short amount of time?? Have you tried selecting the advanced view in BOINC Manager and have you looked on the tasks tab where you could follow all the progress details of each task? Have you checked all the menu items available to you in the Manager to see if you may have inadvertently suspended computation or network activity or something else like this. You don't have any replacement tasks so something is wrongly configured at your end.
At the very least you must have clicked the option not to request more work. If that option wasn't set, you should have got new tasks when your host reported the aborted tasks.
When you start BOINC, have you tried observing all the startup messages in BOINC Manager's event log? It will tell you lots of useful things about what is going on during startup. If you don't understand the messages, you could always try cutting and pasting (into a new post here) about the first 30 or so lines you see immediately after startup. We should be able to see what is going on from that.
Cheers,
Gary.
Hi! The "aborted by user"
)
Hi!
The "aborted by user" message can be misleading, I think in this particular case the tasks actually timed out, since this happened exactly 14 days after the tasks were sent to that host.
But that doesn't change the validity of Gary's advise: for results to be returned successfully to the project, you have to allow BOINC some time to work on them. The question is what keeps your PC from letting BOINC crunch.
Cheers
HB
RE: The "aborted by user"
)
The tasks were aborted around 7 hours after the deadline had expired. There are two things to alert you about expired tasks. Firstly, if you look at the tasks list on the server, expired tasks will have the deadline shown in red (instead of green). If you click on the WU ID link for such a task, you will see whether or not a new task has yet been issued to a new wingman. Even if a new task has been issued to replace the expired one, the expired task is not aborted. It remains there in the quorum with something like "No response" in red. If an expired task is actually completed and reported before the replacement task is reported, it will still count and be awarded credit. At no point will the server mark it as 'aborted' unless it actually is.
The second thing to alert the user is in the startup messages if BOINC is restarted with expired tasks on board. Those messages will warn you and actually tell you that these tasks are "x.y days overdue" and give you the advice that you "should consider aborting" because "you may not receive any credit".
My assumption was that the OP restarted BOINC at some point after the deadline had passed and so noticed the warning and decided to act.
I also wondered if the host in question requested its first tasks and approximately 1 minute after the first one had downloaded completely and started crunching and whilst the second one was still downloading, computation was suspended, and was never at any stage resumed. I think this might even prevent new tasks when the suspended tasks were eventually aborted and reported.
Cheers,
Gary.
RE: RE: The "aborted by
)
Well, yes and no.
Anuj is running a recent BOINC v7.0.28, so we see 'Exit status 203 (0xcb)'. But Einstein is (still) running an ancient version of the BOINC forum code, so the human readable translation of that exit status isn't visible.
I'd like to be able to refer you to [trac]changeset:25601[/trac] and [trac]changeset:25858[/trac], but owing to recent botched changes at BOINC I'll have to refer you to
http://boinc.berkeley.edu/trac/changeset/7f6ab8d49eeaa6df58bf3bc4f61f2f7540d2c7e7/boinc
http://boinc.berkeley.edu/trac/changeset/1f7ddbfe3a27498e7fd2b4f50f3bf9269b7dae25/boinc
instead (Oliver, are you watching?)
Or you could read the original story at
http://boinc.berkeley.edu/dev/forum_thread.php?id=7704
In this case, we're seeing
case 203: return "EXIT_ABORTED_VIA_GUI";so it's likely that the user did abort the two tasks themself.But it is entirely possible that on another occasion we might see
case 200: return "EXIT_UNSTARTED_LATE";when the BOINC client itself aborts a task not started before the deadline, without any human intervention at all. As Crystal Pellet found in July, that would still appear on old message boards (Simap, in that case) as ´Aborted by user´.RE: I'd like to be able to
)
OT: the difference being? A valid comparison would have been to write it like this, which is fully sufficient to uniquely reference the commits: 7f6ab8d and 1f7ddbf
Not that much of a difference I'd say :-) But I agree, Trac sucks.
Oliver
Einstein@Home Project
RE: But it is entirely
)
Yup, confirmed. I have this result that timed out on one of my machines (I did not abort it), it has exit status 200, but YES, it appears as "Aborted by user" on our web pages because of outdated server code.
Thanks for this info!
Cheers
HB
I have finished this task ID
)
I have finished this task ID 322774077 a week ago. Still getting upload backoff after transfer retry.
other tasks had and has no upload and report problem.
I restart that host twice a day (dualboot W7 x Ubuntu), I added :
using sudo gedit /etc/hosts with no success. Tasks is not aborted, database shows it as "in progress".
Any idea?
RE: I have finished this
)
Hmm...why are you adding the IP addresses in the hosts file, don't you have DNS available for the host?
We added additional servers lately, so the list of servers you configured is not complete. To get a complete list (also for other projects you might be running) , you should locate the client_state.xml file in your BOINC directory (the Messages/Event Log view of BOINC Manager has a hint where it is on your system ) and search for instances of "http://". E.g. you missed einstein4.aei.uni-hannover.de) .
Cheers
HB