Downtime
log in

Advanced search

Message boards : News : Downtime

Author Message
Kevin
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 4 Feb 12
Posts: 147
Credit: 2,427,829
RAC: 0
Message 131 - Posted: 13 Aug 2012, 15:46:55 UTC
Last modified: 13 Aug 2012, 15:47:12 UTC

Hi everyone,
We have had some server problems. For this reason, we will not be sending new tasks for a few days.

Sorry for the inconvenience. We will be up and running again soon. New apps and workunits are coming soon too. The new apps should be a bit faster and compatable with a wider range of hosts, while the workunits will be significantly shorter.
Kevin

frankhagen
Send message
Joined: 15 Jul 12
Posts: 48
Credit: 400,146
RAC: 0
Message 132 - Posted: 13 Aug 2012, 16:21:27 UTC - in response to Message 131.

Hi everyone,
We have had some server problems. For this reason, we will not be sending new tasks for a few days.

Sorry for the inconvenience. We will be up and running again soon. New apps and workunits are coming soon too. The new apps should be a bit faster and compatable with a wider range of hosts, while the workunits will be significantly shorter.
Kevin


good news?

i hope you can speed up server side also. with shorter wu's the load will definitely get much higher.

Henk Haneveld
Send message
Joined: 25 Jul 12
Posts: 18
Credit: 78,778
RAC: 0
Message 133 - Posted: 13 Aug 2012, 18:25:47 UTC

The last batch of workunits used on my host on average about 20 minutes runtime. This is compared to work for most other projects short.

Provided you enable checkpointing you should go for longer running workunits not shorter.

If this is not possible you may consider taking a look at what Einstein did. They have a application with short workunits but they bundeled them into a larger result to reduce serverload.

Kevin
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 4 Feb 12
Posts: 147
Credit: 2,427,829
RAC: 0
Message 134 - Posted: 13 Aug 2012, 18:50:04 UTC

Thanks for the feedback.

The workunits are now shorter because we are docking to a smaller region of the protein. We will wait and see how long they take to run and if needed we may bundle them.


As for the server, we are meeting with IT services this week to try and arrange something better.
Thanks,
Kevin

frankhagen
Send message
Joined: 15 Jul 12
Posts: 48
Credit: 400,146
RAC: 0
Message 135 - Posted: 13 Aug 2012, 20:32:42 UTC - in response to Message 134.
Last modified: 13 Aug 2012, 20:51:23 UTC

As for the server, we are meeting with IT services this week to try and arrange something better.


tell them your problem is database performance. if they can't deal with it, oh well - you're toast with shorter wu's..

tng*
Send message
Joined: 9 Aug 12
Posts: 1
Credit: 1,398,473
RAC: 0
Message 136 - Posted: 13 Aug 2012, 23:52:07 UTC - in response to Message 135.

As for the server, we are meeting with IT services this week to try and arrange something better.


tell them your problem is database performance. if they can't deal with it, oh well - you're toast with shorter wu's..


Longer WUs would definitely help the server load, but it shouldn't be a problem to provide the resources for a BOINC project of this size. Speaking as a volunteer, though, I really would like to have longer WUs. WUs as short as the ones you have had are actually rather annoying to me, and the way the BOINC scheduler works seems to shortchange projects with short WUs (they don't sem to run as much as they should based on resource share).

I personally won't care about checkpointing until your WUs start taking a day or so to complete, but I'm sure others will differ with me on that.

Message boards : News : Downtime


Main page · Your account · Message boards


Copyright © 2017 Dr Anthony Chubb