Hi,
since the introduction of the new V0.11 applications, I have been getting a lot of computation errors. Calculation seems alright, but on upload, there is a "filesize too big error".
Thanks
Holger
Reply
Quote
It seems a problem with vv workunits. I found one too:
http://gene.disi.unitn.it/test/result.php?resultid=6650470
Reply
Quote
valtercProject administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 623
Credit: 34,677,535
RAC: 2
It seems a problem with vv workunits. I found one too:
http://gene.disi.unitn.it/test/result.php?resultid=6650470
Ok, my fault, I need to change the "max_nbytes" parameter somewhere in the work generator. I'm just canceling all the vv affected workunits.
Reply
Quote
As of today, I have produced over 50 errors in the last 2 days - they are present for 3 of 4 hosts I'm running. The last one is a Raspberry Pi that has not completed any V0.11 WUs yet.
Checking the units of one host:
http://gene.disi.unitn.it/test/results.php?hostid=4378&offset=0&show_names=0&state=6&appid=
All of the invalid are Vv_vwm tasks. All of the recent valid WUs are Ec_ecm tasks...
Cheers
Holger
____________
My stats
Reply
Quote
valtercProject administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 623
Credit: 34,677,535
RAC: 2
As of today, I have produced over 50 errors in the last 2 days - they are present for 3 of 4 hosts I'm running. The last one is a Raspberry Pi that has not completed any V0.11 WUs yet.
Checking the units of one host:
http://gene.disi.unitn.it/test/results.php?hostid=4378&offset=0&show_names=0&state=6&appid=
All of the invalid are Vv_vwm tasks. All of the recent valid WUs are Ec_ecm tasks...
Cheers
Holger
The Ec workunits were easy to manage: relatively long computation time and small output file, the opposite happens with Vv. Ec (Escherichia coli) has only ~4000 genes while Vv (Vitis Vinifera) more than 28000. So we are trying to balance this in order to achieve the optimal workunit size. (the upload error was obviously not expected, our fault, sorry).
Reply
Quote
Thanks a lot for the information - so the errors are not due to the V0.11 application, but due to WU size on the new project.
BR
____________
My stats
Reply
Quote
valtercProject administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 623
Credit: 34,677,535
RAC: 2
Thanks a lot for the information - so the errors are not due to the V0.11 application, but due to WU size on the new project.
BR
Exactly (the size of the output file)
Reply
Quote
The Ec workunits were easy to manage: relatively long computation time and small output file, the opposite happens with Vv. Ec (Escherichia coli) has only ~4000 genes while Vv (Vitis Vinifera) more than 28000. So we are trying to balance this in order to achieve the optimal workunit size. (the upload error was obviously not expected, our fault, sorry).
Could you make a choice in project preferences to choose the WU type? Some of us (read: 1/2 of the USA) have stone age broadband (a misnomer) speeds and can't handle these large uploads for short runtime WUs. I've had to cut way back because of this.
Reply
Quote