Project has no new tasks available ?!? Really?
log in

Advanced search

Questions and Answers : Web site : Project has no new tasks available ?!? Really?

Author Message
Tim O'Connell
Send message
Joined: 3 Jul 13
Posts: 3
Credit: 29,491
RAC: 0
Message 880 - Posted: 1 Aug 2013, 4:32:01 UTC

Getting a lot of strings like this:

7/31/2013 9:19:20 PM fightmalaria@home Sending scheduler request: To fetch work.
7/31/2013 9:19:20 PM fightmalaria@home Requesting new tasks
7/31/2013 9:19:22 PM fightmalaria@home Scheduler request completed: got 0 new tasks
7/31/2013 9:19:22 PM fightmalaria@home Message from server: Project has no tasks available
7/31/2013 9:19:37 PM fightmalaria@home Sending scheduler request: To fetch work.
7/31/2013 9:19:37 PM fightmalaria@home Requesting new tasks
7/31/2013 9:19:39 PM fightmalaria@home Scheduler request completed: got 0 new tasks
7/31/2013 9:19:39 PM fightmalaria@home Message from server: Project has no tasks available
7/31/2013 9:21:50 PM fightmalaria@home Sending scheduler request: To fetch work.
7/31/2013 9:21:50 PM fightmalaria@home Requesting new tasks
7/31/2013 9:21:52 PM fightmalaria@home Scheduler request completed: got 0 new tasks
7/31/2013 9:21:52 PM fightmalaria@home Message from server: Project has no tasks available

vjahn1
Send message
Joined: 22 Jul 13
Posts: 4
Credit: 271,094
RAC: 0
Message 882 - Posted: 1 Aug 2013, 9:34:39 UTC

Same thing here,

Now the project has no new tasks available. Looks like I'm going to rely on doing
more Malariacontrol WU's instead.

By the way, these problematic issues are getting out of control in that they affect thousands of client machines. Problems such as: "Being out of disk space" is unacceptable. Its easy to forecast disk usage and the choice to delete or archive the data should be routine. The Fightmalaria project should not go down as a result of poor planning or lack of operational control.

So... short of a hardware failure, someone is not managing the projects critical resources well.

Also, lets please don't wait for the user base to tell operations that there is a problem. Operations should be the first to understand an issue and get a quick status report out to the user base for everyone's benefit.

Profile NATE1
Avatar
Send message
Joined: 4 Aug 12
Posts: 37
Credit: 1,065,899
RAC: 0
Message 884 - Posted: 1 Aug 2013, 11:07:03 UTC

unless I'm wrong, I think the admins are new at this and are learning.
so give them time.

mikey
Avatar
Send message
Joined: 21 Jul 13
Posts: 68
Credit: 1,289,610
RAC: 836
Message 886 - Posted: 1 Aug 2013, 13:55:48 UTC - in response to Message 882.

Same thing here,

Now the project has no new tasks available. Looks like I'm going to rely on doing
more Malariacontrol WU's instead.

By the way, these problematic issues are getting out of control in that they affect thousands of client machines. Problems such as: "Being out of disk space" is unacceptable. Its easy to forecast disk usage and the choice to delete or archive the data should be routine. The Fightmalaria project should not go down as a result of poor planning or lack of operational control.

So... short of a hardware failure, someone is not managing the projects critical resources well.

Also, lets please don't wait for the user base to tell operations that there is a problem. Operations should be the first to understand an issue and get a quick status report out to the user base for everyone's benefit.


I can't speak about the problems exactly but normally all that happens automatically in the background with no intervention by the Admins. When the 'disk is full' message shows up it's usually because something else failed, such as the Server handling disk to disk, or server to server, transfers. It also could be the process that failed, not just the server. This is all built into the Boinc Server Side Software, so it should all be automatic. Just the paths need to be input, as a basic explanation.

Kevin
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 4 Feb 12
Posts: 147
Credit: 2,427,829
RAC: 0
Message 887 - Posted: 1 Aug 2013, 18:46:38 UTC

Hi guys,
Sorry about the latest string of issues. The "disk out of space" issue was due to a large number of results being uploaded (> 500,000). These files filled up the disk and as a result the assimilator shut down. We've since changed our assimilator code to make sure it does not switch off in this situation and also freed up some extra space.

The second issue (lots of empty ligands) was related to the first. While there was no space on disk, the work_generator behaved strangely and created tasks with empty input files. We have since modified the work_genertor to fix this issue.

So the root cause of this problem was that we are bottlenecked by our front end server, which is an old desktop computer over 5 years old. We will likely be able to upgrade this soon due to a generous donation. We are bottlenecked by hard disk space and by our ability to handle the increase in traffic on this server as this results in long assimilator queues.

Those of you who have been on this project for a while will be aware that we don't have a constant supply of tasks. There are occasionally interruptions between experiments. This latest interruption is one of these times. We try to keep interruptions to a minimum but they do happen while we prepare the next experiment. This takes time and considerable effort.

Thanks everyone for your feedback,
Kevin

Questions and Answers : Web site : Project has no new tasks available ?!? Really?


Main page · Your account · Message boards


Copyright © 2017 Dr Anthony Chubb