No way to "avoid this d/l problem" - files either do not exist on the server, have a wrong filename, are empty or are in the wrong directory.

Any of this have to be checked and corrected on the server.

Then perhaps its time Radim considered taking a@h offline until he has time to sort out these problems?

Then no one would have to footle about endlessly trying to get WU and failing while the problem got sorted out.

Regards,
Cliff

Among the many "download failed" WUs I do get a few WUs that work. I've even crunched some and sent them back, so I think from my perspective continuing to try is worthwhile. I'm not sure how much work I'm inadvertently causing the Admins, though.

---

800 of my last 940 workunits failed to download correctly. Meanwhile, I'm running the others, with no problems, about 2 hours each as usual. I have 38 waiting for validation. My current queue has 4 running, 4 waiting to run, and 24 download failed.

So, OK, I'm making progress again - with only 11% download-success rate in the past 24 hours.

This looks bad now: 2 tasks running & 32 downloads failed for the same reasons many of us are citing. I aborted all the failed tasks, 32 more downloaded and
*** all failed **
So I have 2 running and it sure looks like *all* downloads are now failing for me.

I will try resetting the project if I see 34 failed downloads & nothing running.

==============

Getting these:

3/20/2015 9:17:36 PM | Asteroids@home | Started download of input_1590_49
3/20/2015 9:17:36 PM | Asteroids@home | [error] File input_1590_53 has wrong size: expected 55719, got 0
3/20/2015 9:17:36 PM | Asteroids@home | [error] Checksum or signature error for input_1590_53
3/20/2015 9:17:36 PM | Asteroids@home | [error] File input_1590_51 has wrong size: expected 55719, got 0
3/20/2015 9:17:36 PM | Asteroids@home | [error] Checksum or signature error for input_1590_51
3/20/2015 9:17:38 PM | Asteroids@home | Incomplete read of 39.000000 < 5KB for input_1590_5 - truncating
3/20/2015 9:17:38 PM | Asteroids@home | Incomplete read of 39.000000 < 5KB for input_1590_49 - truncating

Like everyone else getting lots of download failures. I'm running the following script on my Windows machines. Its basically using BOINCCMD (the command-line tool that comes with BOINC) to hit the update button every 2 minutes. Cut and paste it into Notepad and save as a text file as something like aster.bat and then run it.

echo Script to get/report Asteroids tasks
:1
"C:\Program Files\BOINC\boinccmd" --project http://asteroidsathome.net/boinc update
timeout 120
goto 1

---

BOINC blog

Does anyone recives 'good' WU in the last one hour ?

21/3/2015 09h00

I had been getting 1 or 2 good workunits out of around 50 workunits per update. It is now a 100% failure rate.

I will try resetting the project if I see 34 failed downloads & nothing running.

Resetting the project will not help, don't do it.

---

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

"C:\Program Files\BOINC\boinccmd" --project http://asteroidsathome.net/boinc update

I don't see how this will be of help for the Download problem.

(and just Copy/Paste will not work if boinccmd is not in the default place, 'timeout' is not present on every system)

P.S.
I use similar force-update (times in seconds) but by BoincTasks
(in config.xml in the install directory of BoincTasks):

<config>
	<refresh>
		<downloads>	1709	</downloads>
		<uploads>	1307	</uploads>
		<refresh>	3571	</refresh>
		<refresh_asap>	1399	</refresh_asap>
	</refresh>
</config>

---

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

I'm running 8 machines and have the same problem with downloads.. but enough get through to keep them busy.. however, on this machine, I did 71 WU completed in 24 hours but only have 400 credits.. so validation must be taking awhile, too.. bet Kyong is busy, and I wish I knew how to help.

Yes nearly 99,9999999% fealure rate.

Equal script for linux:

#!/bin/sh
#time in minutes m
SLEEP=2m

while true
do
        sleep $SLEEP
        boinccmd --project http://asteroidsathome.net/boinc/ update
done

Just now my work queue had 30 download failures of 30 tasks. Twice, I aborted all failures & then got 30 more download failures. 90 of 90 for 100% failures. Aborted all again have 1 running & 29 download failures. etc. etc.

Yes, the failure rate is very high. I hope these tasks will be "recycled" & run eventually?

The last three WUs that I tried to download resulted in "Download Error". Guess others are seeing the same thing...

---

same issue here guys, when this will be fixed? ...

Just now my work queue had 30 download failures of 30 tasks. Twice, I aborted all failures & then got 30 more download failures. 90 of 90 for 100% failures. Aborted all again have 1 running & 29 download failures. etc. etc.

Yes, the failure rate is very high. I hope these tasks will be "recycled" & run eventually?

You don't need to abort them, just reporting them is sufficient seeing as they failed to download. Just hit the update button on the project tab or use one of the scripts given earlier in this thread.

They need to fail 7 times before they are flagged as being in error and then the system will stop trying to send them out.

---

BOINC blog

OK, thanks, Mark1. Scripts are fine. I had 1 successful download of the last attempted 560. .0018 is an awfully low download success rate.

Something more strange - a WU file was available (and task done by me) on 17 Mar 2015
but then vanished on 21 Mar 2015 so ended with "Too many errors (may have bug)":
http://asteroidsathome.net/boinc/workunit.php?wuid=27974151

Seems to me something (script? file_deleter?) on the server deletes files too early ...

---

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

Another thing kind of odd... When I looked, I had about 7 or 8 tasks waiting to report. When I did an update to Asteroids in BOINC, and then looked at the events log, it reported that I had 41 completed TASKS instead of the 6, 7, or 8 that actually was sent. Are some of the tasks not reported or delayed? and this is just an outstanding number? When I reported earlier that one of my machines reported 71 tasks in a 24 hour period - that would be impossible, as it can really only do about 10 in 24 hours.

---

I am sorry that there are download errors again. I have checked how the BOINC server is checking the input files that have the same input name as before. The problem is at part files that have already been processed before but not at all of them. So I am going to change naming scheme because of that. New batch of workunits will have different naming scheme so it should finally work without any download errors.

Thanks, Kyong. We ALL appreciate what you do.

---

Should we stop trying to download WUs until later or will it help purge bad WUs if we continue to to try downloading?