Back to normal
log in

Advanced search

Message boards : News : Back to normal

Author Message
Profile Ben
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 17 Nov 14
Posts: 316
Credit: 1
RAC: 0
Message 2070 - Posted: 3 Mar 2015, 10:18:57 UTC

The results database crashed because the server was out of memory...

Some experiments, 18pstiff2re for instance, contained many bad ligand structure. That's why you all had many failed WU.

Plan for the future.
After this batch, I'll wait about one or two weeks in order to get the last remaining WUs. Then, I'll extract all the results, clean the database, clean the server and insert 60 million WU into it. It's a very huge batch so I want to start from a empty database.

Profile Charles Dennett
Avatar
Send message
Joined: 18 Dec 14
Posts: 88
Credit: 3,342,826
RAC: 0
Message 2072 - Posted: 3 Mar 2015, 13:02:50 UTC - in response to Message 2070.

Regarding failed tasks. Some failed with a computation error and some failed with a validation error. Did the bad ligand cause just the computation errors? I saw some indication in the output that might indicate that. All of a sudden, I'm getting no more validation errors. The validation errors all stopped at the same time on all my systems. That makes me think that perhaps your fixing of the results database stopped the validation errors. If that's correct, any chance of revalidating those failed tasks?

Thanks
Charlie
____________

Profile Ben
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 17 Nov 14
Posts: 316
Credit: 1
RAC: 0
Message 2073 - Posted: 3 Mar 2015, 14:42:09 UTC - in response to Message 2072.

Maybe you're right. Bad ligand leads to computation error, not validation.

Profile Ben
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 17 Nov 14
Posts: 316
Credit: 1
RAC: 0
Message 2075 - Posted: 4 Mar 2015, 12:10:02 UTC - in response to Message 2073.

The progress bar will not reach 100%, because of the 400 000 failed tasks.

RNR
Send message
Joined: 12 Apr 13
Posts: 58
Credit: 1,961,436
RAC: 0
Message 2076 - Posted: 5 Mar 2015, 0:25:28 UTC - in response to Message 2075.

Thanks for the information.

Cheers~

Maxxina
Send message
Joined: 11 Dec 13
Posts: 22
Credit: 251,337
RAC: 0
Message 2077 - Posted: 6 Mar 2015, 8:25:55 UTC

So currently we waiting atm for these cca 40 K jobs to return to you . So then you can start clearing database and prepare 60 M WU task ?

Profile Ben
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 17 Nov 14
Posts: 316
Credit: 1
RAC: 0
Message 2078 - Posted: 6 Mar 2015, 10:24:07 UTC - in response to Message 2077.

Yes.

7 days dead line, so except around 2 weeks. I really need to clean everything, the database server will not support the charge. Sorry about that. It also give me some time to do my other project :)

Maxxina
Send message
Joined: 11 Dec 13
Posts: 22
Credit: 251,337
RAC: 0
Message 2080 - Posted: 6 Mar 2015, 11:37:03 UTC

Ok :) There is many projects out there. So will jump to someone :)

Unless i get some new task to crunch . Receiving new data atm. So i got something to do :) 14p2flexre_1 ones. Cool cool :)

Maxxina
Send message
Joined: 11 Dec 13
Posts: 22
Credit: 251,337
RAC: 0
Message 2088 - Posted: 11 Mar 2015, 20:09:53 UTC

Uf, its taking his times, wouldnt be faster to distribute this what is missing between people so they can crunch it fast ? so you could finally close this stuff ? Like yes, people will would have it several times. But this losing time. Dont know finding it little strange

RNR
Send message
Joined: 12 Apr 13
Posts: 58
Credit: 1,961,436
RAC: 0
Message 2089 - Posted: 12 Mar 2015, 0:39:17 UTC - in response to Message 2088.

The desire for completion is inside me as well. Of the 2.9 million there are less than 1,500 WUs left so they'll finish up pretty quickly. But yeah, if they really wanted to send out some duplicates they could finish everything in a few hours ;)

Profile Charles Dennett
Avatar
Send message
Joined: 18 Dec 14
Posts: 88
Credit: 3,342,826
RAC: 0
Message 2090 - Posted: 12 Mar 2015, 12:01:24 UTC

The last few will probably take the longest. I've gotten a few recently and the previous crunchers never did anything with them. They timed out. That's probably the fate of the remaining ones. Maybe if the timeout period could be shortened they might get done quicker.

Also, I don't think you can send a task out to more than one cruncher at a time when all that is needed is one result to be returned.

Charlie
____________

Profile Ben
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 17 Nov 14
Posts: 316
Credit: 1
RAC: 0
Message 2091 - Posted: 12 Mar 2015, 16:54:49 UTC - in response to Message 2090.

I would like to manually reset all the remaining tasks, but I don't know how to do that.

I had a pretty busy week though, did have the time to look at the server.

Maxxina
Send message
Joined: 11 Dec 13
Posts: 22
Credit: 251,337
RAC: 0
Message 2092 - Posted: 18 Mar 2015, 8:41:51 UTC

Well 5 units in last 12 hours or so . With this speed we will be here waiting for another week or so . Cant you move these last units to the next batch ?

Maxxina
Send message
Joined: 11 Dec 13
Posts: 22
Credit: 251,337
RAC: 0
Message 2093 - Posted: 18 Mar 2015, 8:41:59 UTC

Well 5 units in last 12 hours or so . With this speed we will be here waiting for another week or so . Cant you move these last units to the next batch of WU ?

MikeMathis
Send message
Joined: 19 Oct 12
Posts: 4
Credit: 936,099
RAC: 0
Message 2291 - Posted: 26 Apr 2015, 12:26:14 UTC

No work units for the past 15 days
I was successfully running work units until about 4/8 and then nothing.
What can I do to start processing units again?
I have 2 other programs that are working fine.
thanks
Mike

JugNut
Send message
Joined: 8 Nov 12
Posts: 19
Credit: 5,370,953
RAC: 32
Message 2292 - Posted: 26 Apr 2015, 21:10:24 UTC - in response to Message 2291.
Last modified: 26 Apr 2015, 21:23:20 UTC

According to server stats at the time of this post there is no more work available . You'll have to wait until the next batch is released.(supposedly 60mil of them)
http://findah.ucd.ie/server_status.php

If you go by Bens post in this thread that will be 1 or 2 weeks after the last of this batch is completely finished off. Which as per time of this post there are still 263 work units remaining in the wild on other people's PC's that have yet to be crunched & returned so barring the odd resend there's nothing. If you like crunch something else while waiting.
____________

MikeMathis
Send message
Joined: 19 Oct 12
Posts: 4
Credit: 936,099
RAC: 0
Message 2293 - Posted: 27 Apr 2015, 12:56:51 UTC - in response to Message 2292.

thank you much! I _thought_ that the Work Units were done, but not seeing any messages - I wanted to be sure. I'll continue on with my current 2 and then FiND will start up when it can.
Is there any idea of how long that will be?

Profile Ben
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 17 Nov 14
Posts: 316
Credit: 1
RAC: 0
Message 2294 - Posted: 27 Apr 2015, 14:57:45 UTC - in response to Message 2293.

I'm still here. I'm not dead. And I have some code ready.

Vina with the new BOINC API and a work generator with a running time estimation. I just don't have the time to work on uploading everything on the server.

I'll try to find some time to do that :)

MikeMathis
Send message
Joined: 19 Oct 12
Posts: 4
Credit: 936,099
RAC: 0
Message 2303 - Posted: 3 May 2015, 19:07:22 UTC - in response to Message 2294.

Thanks so much Ben. We all are volunteering, but you have the 'heavy lift'. Can I assume that when work is successfully loaded up, my machine will automatically ask for work units?

thanks
Mike

Message boards : News : Back to normal


Main page · Your account · Message boards


Copyright © 2017 Dr Anthony Chubb