Boinc - Scheduler Server Update

[SunRedRX7]

July 8, 2004
We have moved the scheduling server to another machine in order to free up CPU and memory for the DB server. Active clients will fail 10 times to connect to the old scheduling server and then will self correct by contacting the master URL. To force the correction right away, simply detach and then (re)attach the project.

So if you can't connect right now, this is probably why. I just told my machine to update a bunch of times till it was given the new scheduler url.

[JigPu]

'nuther update.

July 8, 2004
Well since yesterday evening, PDT, we have handed out over 600,000 results to be worked on. Thats a six fold increase in overall system performance from our previous best. Now it looks as though we'll have a spotty workunit shortage again as we'll have to bring additional splitters online to keep up with demand. It is 8:50pm PDT and we have about 30,000 results in the queue, the rate we are passing them out is about 1,000 every 10 minutes. Currently the splitters churn out workunits at 4 per minute which creates 12 results per minute.


:eek: So we're chewing through WUs roughly 10x faster than we're creating them?? That's kinda scary.......

JigPu

[Dk Jedi Allianc]

I have made the Detach and attach - and now I can get absolutely NO WORK!?????

[sir_LOIN]

I tried yesterday the detach to try it out and when I merged my two p4 duplicate puters it set my Recent average credit for that one to zero! That was real nice ¦¬(

[Dk Jedi Allianc]

what URL did you attach?????

[SkyHook]

I tried yesterday the detach to try it out and when I merged my two p4 duplicate puters it set my Recent average credit for that one to zero! That was real nice ¦¬(

Yes indeed, as was later pointed out in the Community Forums on the BOINC site, that little tidbit of info about "Detaching" & "Reattaching" was not brought to light when they gave everyone advice about the change of address for the Schedulers. In general be very wary of any advice instructing you to "Detach" or "Reset" the project on your computer unless or until it is absolutely the only way to accomplish something. During the Beta "Detaching" & "Resetting" were quite often the simpliest way of correcting problems that were encountered and didn't do any harm since the work being done wasn't being counted anyway. Now it's a different story and the really dangerous part is not only do you lose the WUs setting waiting to be processed, but also any WUs that have been returned but not yet reported are lost as well since they will now never be reported to the Scheduler.

SkyHook