Hi folks,

The last couple of days I have seen quite a lot of cancelled tasks.

Some examples (there are many more):

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5732279
http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5732095
http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5731457
http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5724675

It looks like tasks are sent out with an extremely short deadline (less than one day). This short deadline expires, an new task is created and sent out and then the original (expired) task reports.

Any clues about what is causing this ?

Tom

Hi folks,

The last couple of days I have seen quite a lot of cancelled tasks.

Some examples (there are many more):

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5732279
http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5732095
http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5731457
http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5724675

It looks like tasks are sent out with an extremely short deadline (less than one day). This short deadline expires, an new task is created and sent out and then the original (expired) task reports.

That isn't a deadline, that is the time the Wu was reported after being cancelled (the column is dual use: Time reported or deadline)

Any clues about what is causing this ?

Eric increased the initial replication to three:

http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2046&postid=46724

I've increased the minimum replication to 3 and reduced the per CPU/GPU max in-progress to 50. It does mean fast hosts will need to contact the server more than once a day, but it should improve the validation speed for "rare" apps like Android to the point where I can make more sense of it.

Once two results are in, the third task is cancelled as it is no longer needed, if you don't want to get cancellations then set a minimal cache and start your Wu's promptly.

Claggy

Hi Claggy,

Thanks for the explanation.
I should have spotted the 'initial replication' on the workunit page.

Tom

-[ snip ]-

-[ snip ]-

I've increased the minimum replication to 3 and reduced the per CPU/GPU max in-progress to 50. It does mean fast hosts will need to contact the server more than once a day, but it should improve the validation speed for "rare" apps like Android to the point where I can make more sense of it.

Once two results are in, the third task is cancelled as it is no longer needed, if you don't want to get cancellations then set a minimal cache and start your Wu's promptly.

Claggy

Greetings Claggy,

Thanks for this. I was about to question this myself when I found this thread.

I just could not understand why the work unit page said "Aborted by user" when I do not abort WUs. Thanks again. :)

Keep on BOINCing...! :)

---

CAPT Siran d'Vel'nahr XO
USS Vre'kasht NCC-33187

Siran's website: [ ONLINE! ]