Been doing anything but deferring communication for a certain period of time, but yet others seem to have no problem getting work from BOINC.

> Been doing anything but deferring communication for a certain period of time,
> but yet others seem to have no problem getting work from BOINC.
>

I don't really know what problem you are having, but I found that if your do not force feed your reporting, it usually goes well. I have tried to update the project when I did not have WU's to transfer, and I constantly get "Deferring Communication". But when I am sending WU"s in, they take, I have been getting 2 to 3 new WU's in the same process. I disable communication to the network for 3 to six hours at a time. When I get back online I renable the network, and off she runs.
I don't know if this will help, but I hope it does.

Regards,


Rocky

I have hit the update button a couple of times but it was saying
"no work from project" untill I finally recived 2 wu. I kept pushing
the update button a few times and again "no work from project" until
I received 10 in a row.

To me no work from project mean no work from project witch is
apparently not the case...I guess they should change the no work
from project message by "No luck try again"

> I have hit the update button a couple of times but it was saying
> "no work from project" untill I finally recived 2 wu. I kept pushing
> the update button a few times and again "no work from project" until
> I received 10 in a row.
>
> To me no work from project mean no work from project witch is
> apparently not the case...I guess they should change the no work
> from project message by "No luck try again"
>
>
I think no work is still the proper message. The no work status can be transitory though since when all is going smoothly the splitter/feeder will be constantly generating more work.

John Keck
BOINCing since 2002/12/08

I get at various times:

1. No Work From Project
2. No Work Available
3. Not-Sending Work - Last RPC too recent: 8 sec
4. Daily Quota Exceeded
5. No schedulers responded
6. And all sorts of other error messages thrown in to boot ...

And then Booooom 10 to 20 WU's downloaded.... :)
I quit believing anything the BOINC GUI tells me a long time ago ... ;)

Unless you have a REALLY fast computer or are downloading a monster sized work queue, then you souldn't get the Daily Quota Exceeded message unless there is a problem. The daily limit per machine is 50. From what I've seen, it uses the number downloaded since midnight (UCB Time, your time, or GMT I don't know), which would mean that at the time you see that, the machine giving you that message has downloaded 50 workunits since midnight. You might want to check that one out.

Chris



> I get at various times:
>
> 1. No Work From Project
> 2. No Work Available
> 3. Not-Sending Work - Last RPC too recent: 8 sec
> 4. Daily Quota Exceeded
> 5. No schedulers responded
> 6. And all sorts of other error messages thrown in to boot ...
>
> And then Booooom 10 to 20 WU's downloaded.... :)
> I quit believing anything the BOINC GUI tells me a long time ago ... ;)
>

I keep getting these messages and then it just hangs at the last message:

SETI@home - 2004-08-06 07:38:19 - Project prefs: using your defaults
SETI@home - 2004-08-06 08:22:20 - Sending request to scheduler: http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi
SETI@home - 2004-08-06 08:22:24 - Scheduler RPC to http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi succeeded

I did get it to download one WU in the last 24 hours by hitting the update button.

Anybody elsel getting this and if so any suggestions? I don't know if I should keep hitting the update button what with the problems they are having.

Thanks and have a great day!!

James

http://www.palantir.net/cgi-bin/file.cgi?file=wav/foolprf.wav

I've yet to figure out the magic keystokes or mouseclicks. But one of my machines is happily cruching away at seti/boinc with units downloaded last nite, while the other has been without for what seems like going on the third week now. Both machines have been set up the same as near as I can tell, but one is consistantly without crunchies.

Fortunately both keep busy when there are no BOINC units available, SETI/Classic gets powered up on those lean days.

I had similar problems yesterday they seem to have been cleared up by down loading ver 3.2 of boinc, I was running 3.19 some how missed the announce ment that a newer verison was avail. All my boxes have work today.

To check your version in bonic clik help tab ,then click about, the version will be displaied.
Hope this helps.

> I get at various times:
>
> 1. No Work From Project

Either due to not meeting requirements (like disk-usage), but often instead of "no work available".

> 2. No Work Available

Should be self-explanatory, they currently haven't any work to give out. But since they're constantly generating new work, the next asking for work can get some...

> 3. Not-Sending Work - Last RPC too recent: 8 sec

You can only talk to the scheduler once every 10 minutes. Often happens automatically if couldn't fill your cache in last asking.

> 4. Daily Quota Exceeded

You've got assigned 50 wu since midnight Berkeley-time, or 07:00 UTC. If you're got fast machine & 10+ days cache-setting, can take over 3 days to fill your cache. But if you're running an old p2 or similar, look for download-errors or client-errors.

> 5. No schedulers responded

Down for maintanance, currently being swamped with requests so won't talk to you, network-problems, you're swamping your line with other traffic so BOINC-connection times out.

> 6. And all sorts of other error messages thrown in to boot ...

Not seen any, except wrong major version, or too old core client you need to upgrade...

BTW, the whole point of exponential backoff is "to avoid flooding servers".
Constantly hitting "retry" isn't helping things out, and atleast for uploading results doesn't even make sense if you've got a permanent connection, since the max retry-time is 4 hours.
Of course with wrongly showing errors like "you must use major version 4" trying again before 24 hours is expired is another thing. The same if you detects it's your line that dropped out or wrong proxy-settings. ;)

> I get at various times:
>
> 1. No Work From Project
> 2. No Work Available
> 3. Not-Sending Work - Last RPC too recent: 8 sec
> 4. Daily Quota Exceeded
> 5. No schedulers responded
> 6. And all sorts of other error messages thrown in to boot ...
>
> And then Booooom 10 to 20 WU's downloaded.... :)
> I quit believing anything the BOINC GUI tells me a long time ago ... ;)
==========

EErrrrmmmm....What I was getting at was I can get all those messages within a Couple of Minutes time frame...So If there was no work available 1 minute and then I download 20 WU's the next minute why did I get the No Work Available in the First place if indeed there was work available...

All I was trying to do was point out the Messages you can get & why I quit believing what their telling me because I know better, I have no problem getting work if I want it & there is some available no matter what the Server tells me, it's been telling me since yesterday that there is no work available but from somewhere I've managed to download around 300 Wu's onto my PC's... No work available my arss...;)

My second machine is finally downloading units..even tho it is only 2 at a time. At least 4 to crunch over the weekend. Maybe more as time goes on.

Hooray!!!

>
> EErrrrmmmm....What I was getting at was I can get all those messages within a
> Couple of Minutes time frame...So If there was no work available 1 minute and
> then I download 20 WU's the next minute why did I get the No Work Available in
> the First place if indeed there was work available...
>

Let's say the transitioner due to I/O-limits adds 10 "results" every 2 seconds in a block, while the scheduler can service one machine/second. Since scheduler max can give 20 results/asking, it's not difficult to understand the system can be out of work 1 second but have work available the next. ;)

Of course seti for BOINC is different, and since the bottleneck is the database the adding of "results" is probably more random so it's little help decreasing the max the scheduler can give out. Also, asking 50 times and getting 1 result/asking is maybe more taxing than giving 50 wu in 3 askings.

Oh, and maybe to confuse people they're instead saying "no work available" on even seconds, while "no scheduler responded" on odd, and only accept connections every 10 seconds. :)