Версія даної теми для друку

Автор: spingadus Sep 15 2012, 17:37

I just had 8 long tasks permanently fail on upload. All tasks completed. Is this issue known to the project team?

Could I please get some credit for the tasks?

http://dg.imp.kiev.ua/slinca/results.php?hostid=2101&offset=0&show_names=0&state=5


Errors example (using Boinctasks gui):

9918 SLinCA@Home 9/15/2012 7:45:08 AM Computation for task 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0 finished
9919 SLinCA@Home 9/15/2012 7:45:10 AM Started upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_0
9920 SLinCA@Home 9/15/2012 7:45:10 AM Started upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_1
9921 SLinCA@Home 9/15/2012 7:45:10 AM Started upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_2
9922 SLinCA@Home 9/15/2012 7:45:10 AM Started upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_3
9923 SLinCA@Home 9/15/2012 7:45:13 AM [error] Error reported by file upload server: invalid signature
9924 SLinCA@Home 9/15/2012 7:45:13 AM [error] Error reported by file upload server: invalid signature
9925 SLinCA@Home 9/15/2012 7:45:13 AM [error] Error reported by file upload server: invalid signature
9926 SLinCA@Home 9/15/2012 7:45:13 AM [error] Error reported by file upload server: invalid signature
9927 SLinCA@Home 9/15/2012 7:45:13 AM Giving up on upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_0: permanent upload error
9928 SLinCA@Home 9/15/2012 7:45:13 AM Giving up on upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_1: permanent upload error
9929 SLinCA@Home 9/15/2012 7:45:13 AM Giving up on upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_2: permanent upload error
9930 SLinCA@Home 9/15/2012 7:45:13 AM Giving up on upload of 4e9d0ba6-7c7a-4810-9fe2-eb5ffc0e41b8_b1191264-dc90-4e00-9382-abb086425454_471_0_3: permanent upload error

Автор: AMDave Oct 8 2012, 12:25

me too now.
[error] Error reported by file upload server: invalid signature


client versions x.34, x.36, x.38
Linux x86_64 and Win 7 x86_64
boinc 7.0.28 and 7.0.31 (which were all working up until about 1 week ago)

I have to give up for a while. No point in
This is going nowhere until admins have time to spend fixing the problems and test them and tell us they have done so.
Plenty of others to crunch while we wait

Автор: AMDave Oct 10 2012, 05:41

32.32 also failed.

Re-trying with clean install OS linux x86_64 + ia32libs + BOINC 7.0.28
got a 32.38 WU.
will post result.

Автор: Rilian Oct 10 2012, 06:44

hi AMDave, !

I have emailed to YuRi via gord [A] imp.kiev.ua but he did not reply yet!

Unfortunately i cannot help with this error

Автор: AMDave Oct 10 2012, 09:21

Thanks Rilian.
It may be important to note there are many ISP & Carrier upgrade and maintenance changes occurring on the internet at the moment.
Getting very bad connections speeds from AU to other countries, (ping varying from 41ms to 5000ms) & lots of intermittent DNS lookup failures.
These things should not relate to sig 11 errors on the WUs unless there is a "trickle update" that fails.
(I am not aware if SlinCA has a trickle update. I don't remember seeing one in the project)

As per earlier, I retested with clean installs and still get 100% fail rate. win7_AMD64 and linux_AMD64 with ia32 libs installed.
Other projects working but getting unusual increase in failed WUs on some of them also, so I thought I'd mention the internet issues.

YuRi & gorg may be able to look at the server stats and tell us if this problem is just me, or if others are having an increase in errors as well.

BTW. You have helped me with this error by letting the right people know. Many Thanks!

Автор: AMDave Oct 11 2012, 07:59

connectivity & DNS issues are less frequent now.
Trying again.
I have a 32.34 wu in progress on linux AMD64 server, at 7hrs 47 mins now.
see how it goes

Автор: AMDave Oct 11 2012, 22:29

It is almost done. 2hrs left.
It seems like the Signal 11 errors were caused by the internet issues.

Автор: AMDave Oct 12 2012, 00:11

Aw NUTZ!!

Run time 84895.62

<core_client_version>7.0.28</core_client_version>
<![CDATA[
<message>
process exited with code 1 (0x1, -255)
</message>
<stderr_txt>
sh: 1: zip: not found
Could not open ZIP_local file.
07:45:52 (3089): called boinc_finish

</stderr_txt>
]]>

Oh well, at least I'm back to the known issues. ha ha.

Автор: AMDave Oct 12 2012, 00:21

I re-installed zip
verified it is at /usr/bin/zip

I have some more wu's in progress to test again

Автор: Rilian Oct 12 2012, 10:20

Hi Dave, still no reply from Yuriy

i will try to reach him in another way

thank you for the patience! I do not crunch this project so i cannot help, unfortunately

Автор: AMDave Oct 12 2012, 10:55

Thanks Rilian, but there is no more need.

I have just returned some WUs successfully.

It seems that the SIG 11 errors were caused by the BOINC client itself during the carrier DNS & proxy cache problems.
Now that is solved, I moved the client to a LInux server which did not have the "zip" alias installed.
Once I installed the "zip" alias, all seems to be working again now.

Please send them a follow up message saying it is all ok now. No need for concern.
Thanks again.

Автор: Rilian Oct 12 2012, 15:12

hi AMDave, great success!!!!!

could you please give an advice (in short) how you have fixed this, we will ask Slinca admins to put this on the project news or somewhere

as i understand, Slinca core is statically linked to some ZIP path like /usr/bin/zip and you had another path in your system ?

Автор: spingadus Oct 18 2012, 10:57

So, basically this was a zip executable path issue? Was there any specific instructions to follow in order to run tasks on my win7-64 machine? I have installed 7zip. Could that have caused a problem. Sorry if this is a newbie question.

I'm going to try running this on linux either way.

Автор: AMDave Oct 19 2012, 13:44

I do not believe so. No.
After that WU was returned successfully, more WUs failed.
I am unable to replicate a set of circumstances where a WU either fails or succeeds.
Until a few hours ago every WU since then had failed.
Other projects continue successfully on the same boxen.
I have just returned some more successful work units after numerous failures on the same platform - without changes.
There does not seem to be any consistent reason for a WU to fail or succeed - that I can tell.
I will continue until I return 1 more successful WU to get to the 100K milestone and then I will take a break from this project for a while.

Автор: spingadus Oct 19 2012, 20:06

Thanks for the reply AMDave. I'm currently running only 2 tasks in an ubuntu64 box. Let's hope that they don't fail as well. I wish the tasks weren't so long. It's painful to tie up resources for so long only to have them fail.

Автор: skgiven Nov 12 2012, 13:19

I ran 4 task. All failed after about 17.30h with 'Computation Error'. They might have failed on completion.
Since Saturday, I have not been able to upload tasks.
The server was down yesterday, so I aborted any task that were queued.

The server status page still times out,
dg.imp.kiev.ua/slinca/server_status.php

Я побіг 4 завдання. Всі вдалося приблизно через 17.30h з розрахунком помилки. Вони, можливо, не вдалося на доопрацювання.
З суботи, я не був в змозі завантажити завдань.
Сервер впав вчора, так що я перервати будь-які завдання, які були поставлені в чергу.

На сторінці стану сервера ще раз з,
dg.imp.kiev.ua / SLinCA / server_status.php

Я пропоную вам додати модуль Google Translator у свій сайт.