Blurb (Jun 26 2007)

Regular outage today for backup compression/backup. Bob took care of all that. Meanwhile I briefly shut down the Network Appliance to clean its pipes, pull out some bad drives, and re-route some cables. This caused the web servers to all hang for about 5 minutes. Other than that, just playing with the new toys. New multibeam client probably won't be happening this week. Eric is still digging himself out from some long days of grant proposal writing last week.

- Matt

____________
-- BOINC/SETI@home network/web/science/development person
-- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude

Matt , thank you very much for the update !

Kind Regards
Byron

Eric is still digging himself out from some long days of grant proposal writing last week.

Thanks for the UIpdate...

btw nor Probs here in Germany

Greetings from Germany NRW
Ulli

The only thing I would ask is that you give us least 24 hours advance notice before rolling MB out so we can take appropriate steps. Especially for the folks (like myself) who run the Coop and other apps using the anonymous platform.

Alinator

MB roll out: in addition to some leading notice, would someone also explain expected behavior. Do we need to download new clients, for example? Will we be wise to adjust our caches? Should we turn to other projects for a couple of weeks to minimize cpu cycle loss due to chaos? You know, the usual concerns.

MB roll out: in addition to some leading notice, would someone also explain expected behavior. Do we need to download new clients, for example? Will we be wise to adjust our caches? Should we turn to other projects for a couple of weeks to minimize cpu cycle loss due to chaos? You know, the usual concerns.

We saw this with the change from 4.x to 5.x.

For those running standard science applications (no app_info.xml), BOINC will download the new science application when it receives the first 5.2x (multibeam) work units. It will be completely automatic.

If you have an app_info.xml, and it does not specify anything for 5.2x (it should only specify versions that are known to work and not some future unreleased version that might or might not be okay), then the BOINC servers will only send 5.15.

For a while, there will be 5.15 and 5.2x units to be crunched (replacements for lost work, no quorum, etc.) and at first, there will be plenty.

As the 5.15 work gets scarce, those with an outdated science application and app_info.xml will run out of work and go idle.

In other words, this isn't a big deal. You aren't going to have days or weeks of "idle time" unless you just plain miss it -- and since most of the people asking for advance notice watch BOINC closely, I don't see that happening.

If you are really worried about the transition, delete your app_info.xml now, let BOINC revert to the standard science application, and let it manage the transition -- then update to the new chicken app. that handles multibeam.

Note that there is some need to keep both 5.15 and 5.2x "out there" or some of the earlier work units will never form a quorum and never get credit. This will likely get handled by those running the standard, non-optimized science app.

So it sounds like if I do nothing, then I will not see a difference at first, but in time, as the 5.15 wu's run out I will see more MB wu's and a general decrease in RAC. This latter effect will be remedied by moving to an MB updated application. But if the reduction in RAC was not important to me, then I could continue running my optimized app for the foreseable future, without any changes.

Right?

Red = generalizing from a different situation, so off-track.