All GPU jobs which has finished, get's an error uploading/downloading the job. Status in BoincTasks 1.61 reports "Download Error"

Your posted log show only Download errors (so these can't be tasks "which has finished").
Do you also have uploading errors in the log?

(I don't have Upload errors, I only have this: (which is during Reporting, not Uploading)

2673 Asteroids@home 30-09-2014 10:03 Sending scheduler request: Requested by project.
2674 Asteroids@home 30-09-2014 10:03 Reporting 1 completed tasks, not requesting new tasks
2675 Asteroids@home 30-09-2014 10:04 Scheduler request completed
2676 Asteroids@home 30-09-2014 10:04 Message from server: Server can't open database
)

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

 
The other problem with "Error while downloading" is that some tasks
which are already completed by someone
can't be done by anyone else so the WU ends with:
"Completed, can't validate"/"Too many errors (may have bug)"
http://asteroidsathome.net/boinc/workunit.php?wuid=20964261


As this is bad for the project (some WUs have no chance to generate results)
I sent PM to Kyong about the problem

(definition: WU = collection/sum of all tasks, in this case 10 tasks were sent/generated for this WU (but 8 were not received))
 

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

Some more WUs in this state (most from users which posted in this thread)

All have one task "Completed, can't validate"
and the WU ended in "Too many errors (may have bug)"

http://asteroidsathome.net/boinc/workunit.php?wuid=20957669
http://asteroidsathome.net/boinc/workunit.php?wuid=20908295
http://asteroidsathome.net/boinc/workunit.php?wuid=20894908

http://asteroidsathome.net/boinc/workunit.php?wuid=20894883
http://asteroidsathome.net/boinc/workunit.php?wuid=20966854
http://asteroidsathome.net/boinc/workunit.php?wuid=20968662

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

The problem appears to have been corrected. I'm 24 hours without an error...

They are back again

<core_client_version>7.4.21</core_client_version>
<![CDATA[
<message>
WU download error: couldn't get input files:
<file_xfer_error>
<file_name>input_304409_7</file_name>
<error_code>-224 (permanent HTTP error)</error_code>
<error_message>permanent HTTP error</error_message>
</file_xfer_error>

</message>
]]>

this is one of three failed downloads.


Edit: while I waited for the update of this post 2 more download errors occured.

FOUR OF FOUR ATTEMPTED DOWNLOADS FAIL IMMEDIATELY:

Asteroids@home 10-10-2014 09:49 work fetch resumed by user
Asteroids@home 10-10-2014 09:49 Sending scheduler request: Requested by project.
Asteroids@home 10-10-2014 09:49 Requesting new tasks for CPU
Asteroids@home 10-10-2014 09:49 Scheduler request completed: got 1 new tasks
Asteroids@home 10-10-2014 09:49 Started download of input_312508_2
Asteroids@home 10-10-2014 09:49 Giving up on download of input_312508_2: permanent HTTP error

Stderr output
<core_client_version>7.0.36</core_client_version>
<![CDATA[
<message>
WU download error: couldn't get input files:
<file_xfer_error>
<file_name>input_312508_2</file_name>
<error_code>-224</error_code>
<error_message>permanent HTTP error</error_message>
</file_xfer_error>

</message>
]]>

.......... etc

Asteroids@home 10-10-2014 09:50 Sending scheduler request: To report completed tasks.
Asteroids@home 10-10-2014 09:50 Reporting 1 completed tasks
Asteroids@home 10-10-2014 09:50 Requesting new tasks for CPU
Asteroids@home 10-10-2014 09:50 Scheduler request completed: got 1 new tasks
Asteroids@home 10-10-2014 09:50 Started download of input_313094_9
Asteroids@home 10-10-2014 09:50 Giving up on download of input_313094_9: permanent HTTP error

Asteroids@home 10-10-2014 09:52 Sending scheduler request: To report completed tasks.
Asteroids@home 10-10-2014 09:52 Reporting 1 completed tasks
Asteroids@home 10-10-2014 09:52 Requesting new tasks for CPU
Asteroids@home 10-10-2014 09:52 Scheduler request completed: got 1 new tasks
Asteroids@home 10-10-2014 09:52 Started download of input_312626_2
Asteroids@home 10-10-2014 09:52 Giving up on download of input_312626_2: permanent HTTP error

Asteroids@home 10-10-2014 09:53 update requested by user
Asteroids@home 10-10-2014 09:53 Sending scheduler request: Requested by user.
Asteroids@home 10-10-2014 09:53 Reporting 1 completed tasks
Asteroids@home 10-10-2014 09:53 Requesting new tasks for CPU
Asteroids@home 10-10-2014 09:53 Scheduler request completed: got 1 new tasks
Asteroids@home 10-10-2014 09:53 Started download of input_312580_1
Asteroids@home 10-10-2014 09:53 Giving up on download of input_312580_1: permanent HTTP error

---

Same here, 2 out of 4 failed:

10.10.2014 17:48:19 | Asteroids@home | Sending scheduler request: Requested by user.
10.10.2014 17:48:19 | Asteroids@home | Requesting new tasks for CPU and NVIDIA GPU
10.10.2014 17:48:21 | Asteroids@home | Scheduler request completed: got 4 new tasks
10.10.2014 17:48:23 | Asteroids@home | Started download of input_313394_2
10.10.2014 17:48:23 | Asteroids@home | Started download of input_313148_4
10.10.2014 17:48:24 | Asteroids@home | Finished download of input_313394_2
10.10.2014 17:48:24 | Asteroids@home | Giving up on download of input_313148_4: permanent HTTP error
10.10.2014 17:48:24 | Asteroids@home | Started download of input_313396_3
10.10.2014 17:48:24 | Asteroids@home | Started download of input_311399_6
10.10.2014 17:48:25 | Asteroids@home | Finished download of input_313396_3
10.10.2014 17:48:25 | Asteroids@home | Giving up on download of input_311399_6: permanent HTTP error


I had this sometimes recently but didn't know this was a more serious error until I looked up the message boards now.

Please check. Thanks!
:)

Still failing and getting worse: 6 out of 8 failed.

Both SSE3 and CUDA55 WUs are affected.

Both SSE3 and CUDA55 WUs are affected.

There is no difference between "SSE3 and CUDA55 WUs", so if this happens it will happen with any 'type' of WU
(e.g. the same task you receive for SSE3 app someone else will receive for CUDA or AVX or ...)

Some of these 'Error while downloading' is yours, the other 8 are your 'wingmates' (quorum partners) - the same WU, 9 tasks:
http://asteroidsathome.net/boinc/workunit.php?wuid=21710924

Really bad is when somebody (me in the case ;) )
Completed the task but there is no other wingmate that can confirm the result
so the final outcome is 'Completed, can't validate' (after 9 Error tasks)
http://asteroidsathome.net/boinc/workunit.php?wuid=20964261
 

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

Both SSE3 and CUDA55 WUs are affected.

There is no difference between "SSE3 and CUDA55 WUs", so if this happens it will happen with any 'type' of WU
(e.g. the same task you receive for SSE3 app someone else will receive for CUDA or AVX or ...)
 

Pardon me, but SSE3 WUs are for CPU, while CUDE55 WUs are for GPU.
But I guess you meant that there is no difference in WU generation, right?

I just thought I should mention that both types are affected, so the problem can be found more easily.

Some of these 'Error while downloading' is yours, the other 8 are your 'wingmates' (quorum partners) - the same WU, 9 tasks:
http://asteroidsathome.net/boinc/workunit.php?wuid=21710924

Really bad is when somebody (me in the case ;) )
Completed the task but there is no other wingmate that can confirm the result
so the final outcome is 'Completed, can't validate' (after 9 Error tasks)
http://asteroidsathome.net/boinc/workunit.php?wuid=20964261

I didn't have that myself (yet), but sure, that's quite annoying.
But aren't tasks with download errors be sent out again automatically?
If not, why not?

Houston, we have a problem.

I'm fresh out of tasks, each request for work comes back empty...
10/12/2014 11:13:28 PM | Asteroids@home | Scheduler request completed: got 0 new tasks
10/12/2014 11:13:28 PM | Asteroids@home | Project has no tasks available

...and there are only 19 tasks ready to send.

Crunching in Clermont.
Carl

Pardon me, but SSE3 WUs are for CPU, while CUDE55 WUs are for GPU.
But I guess you meant that there is no difference in WU generation, right?

I mean that the same file you try to Download from the server 'for CUDA' (= to be processed by CUDA application), someone else will try to Download for CPU
So there is no difference in the WU file (if the file do not exist on the server = Download errors for everyone)

The only difference is that you will process the Downloaded file by CUDA application,
the other computer will use other application (for CPU or again for CUDA - depend what his BOINC asked for)

But aren't tasks with download errors be sent out again automatically?
If not, why not?

Click the links and see - the WU was "sent out again automatically" 10 times and failed on every attempt
(and the server gave up: 'Too many errors (may have bug)' - no more attempts will be made)
 

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

Houston, we have a problem.

I'm fresh out of tasks, each request for work comes back empty...
10/12/2014 11:13:28 PM | Asteroids@home | Scheduler request completed: got 0 new tasks
10/12/2014 11:13:28 PM | Asteroids@home | Project has no tasks available

...and there are only 19 tasks ready to send.

Crunching in Clermont.
Carl

The webpage is cached, NOT a live page, so the "19" you see may not be a real number, just what was available the last time the page was updated. In your case it seems to in fact be zero. In fact the page now shows:
Tasks ready to send 0

Hi Everyone,

New to this! I am also getting the download errors. 19 of previous 60 tasks have been download errors.

Gary

Yes, I've seen > 30% download failures for several days. Some fail immediately. And no tasks available for about 1 day. Server status are OK, but there's no work available.

Pardon me, but SSE3 WUs are for CPU, while CUDE55 WUs are for GPU.
But I guess you meant that there is no difference in WU generation, right?

I mean that the same file you try to Download from the server 'for CUDA' (= to be processed by CUDA application), someone else will try to Download for CPU
So there is no difference in the WU file (if the file do not exist on the server = Download errors for everyone)

The only difference is that you will process the Downloaded file by CUDA application,
the other computer will use other application (for CPU or again for CUDA - depend what his BOINC asked for)

I wasn't aware of that but thought that SSE jobs are always sent out as SSE and CUDA as CUDA. On the other hand, one could have guessed that it makes no difference so they get mixed up. Thanks for the info!
:)

But aren't tasks with download errors be sent out again automatically?
If not, why not?

Click the links and see - the WU was "sent out again automatically" 10 times and failed on every attempt
(and the server gave up: 'Too many errors (may have bug)' - no more attempts will be made)
 

I see. Thanks again!

They are back again!
Isn't there a vaccine available against this?? :))

Yep, I've got 15 download errors already today (for 2 different computers). Total 173 on the list... from the recent days. Also there seems to be daily problems to upload results of completed tasks. Upload typically needs numerous retrys and takes time.

I have download errors mostly checksum or signature errors

They are back again!
Isn't there a vaccine available against this?? :))

Yes shorter internet travel from the Server to you, ie fewer hops!