Do you have any way of easily checking if the upload files for the task he originally mentioned are still sitting on the upload server?
Or put another way, what files should i expect to see if they had not been uploaded? I do recall the status being 100% uploaded but remember nothing in the job_log_* for this task. (hence the - i expect the invalid result)
Christian Beer wrote:
Also in your latest post where you add two lines you seem to add them to a O1Spot1 task and not the stuck FGRPB1G task was this just a copy and paste error?
Yes I had another window open at the time on a different host to look at the layout of the file. Apologies for the confusion there.
Do you have any way of easily checking if the upload files for the task he originally mentioned are still sitting on the upload server?
Or put another way, what files should i expect to see if they had not been uploaded? I do recall the status being 100% uploaded but remember nothing in the job_log_* for this task. (hence the - i expect the invalid result
The validator directly operates on the upload server so if it can't find the files they are not there. I also checked that this is really the case. The local files are in the project directory and have the same name as the task wit an added _0 and _1 at the end. I don't think you can see if they are uploaded or not as they usually only get deleted after the task was reported. So you can't really check if the files were uploaded or not.
The local files are in the project directory and have the same name as the task wit an added _0 and _1 at the end. I don't think you can see if they are uploaded or not as they usually only get deleted after the task was reported. So you can't really check if the files were uploaded or not.
Thanks Christian, I do recall looking for files matching the task name at the time, hence the comments about the template files.
Some more digging reveals - on this host around the time.
between
Wed 14 Jun 01:07:51 BST 2017 and Wed 14 Jun 01:32:22 BST 2017
the job_log file has a 116 null characters added in place of a record.
and "# last reboot" reveals a restart occurred at 01:18
which looks a smoking gun.
I would suggest the output files were either empty and not closed properly or some such. I'm not at all worried about losing the task.
The only thing is, boinc should be fault tolerant and know what to do in these cases, and have some graceful failure / resend the upload / abort task instead being stuck.
I had this problem with a couple of WCG tasks, they were preventing uploads of other good tasks so had to keep 'aborting' them every day to clear the backlog, a bit of googling brought me here and this fixed them perfectly. So a big thank you to all the contributors on this thread.
Gary Roberts wrote:Do you
)
Or put another way, what files should i expect to see if they had not been uploaded? I do recall the status being 100% uploaded but remember nothing in the job_log_* for this task. (hence the - i expect the invalid result)
Yes I had another window open at the time on a different host to look at the layout of the file. Apologies for the confusion there.
AgentB wrote:Gary Roberts
)
The validator directly operates on the upload server so if it can't find the files they are not there. I also checked that this is really the case. The local files are in the project directory and have the same name as the task wit an added _0 and _1 at the end. I don't think you can see if they are uploaded or not as they usually only get deleted after the task was reported. So you can't really check if the files were uploaded or not.
Christian Beer wrote:The
)
Thanks Christian, I do recall looking for files matching the task name at the time, hence the comments about the template files.
Some more digging reveals - on this host around the time.
between
Wed 14 Jun 01:07:51 BST 2017 and Wed 14 Jun 01:32:22 BST 2017
the job_log file has a 116 null characters added in place of a record.
and "# last reboot" reveals a restart occurred at 01:18
which looks a smoking gun.
I would suggest the output files were either empty and not closed properly or some such. I'm not at all worried about losing the task.
The only thing is, boinc should be fault tolerant and know what to do in these cases, and have some graceful failure / resend the upload / abort task instead being stuck.
I had this problem with a
)
I had this problem with a couple of WCG tasks, they were preventing uploads of other good tasks so had to keep 'aborting' them every day to clear the backlog, a bit of googling brought me here and this fixed them perfectly. So a big thank you to all the contributors on this thread.