|Home||Reviews||Tools||Forums||FAQs||Find Service||ISP News||Maps||About|
how-to block ads
Upon transmission of the RtR to the Scheduler (a different part of the server system processes), a brief 'Acknowledgement' will be seen in the Tasks tab Status column, after which the Task entry in the client is cleared and all traces in the local tracking files, except for the Message Logs is removed.
For purposes of efficiency, the Ready To Report is send on the next 'Scheduled' contact with the servers and are per the below table.
New work is needed. Completed work will be reported at the same time.
Sample message log entries:
15/12/2007 10:29:23|WCG|Sending scheduler request: To fetch work. Requesting 39440 seconds of work, reporting 6 completed tasks15/12/2007 10:29:28|WCG|Scheduler request succeeded: got 4 new tasks
The Deadline is less than 24 hours away.
The Result is less than the Work Buffer setting from it's Deadline (New for v5.4.xx)
The Result has been in the Ready to Report status for 24 hours. (New for v5.10.14)
The Result has been in the Ready to Report status longer than the Work Buffer setting. (Removed in v5.10.14 and later)
If "net_status.have_sporadic_connection" will report immediately. (possibly for dialup-users only!)
If a 'trickle' is uploaded to the server (not applicable to WCG and only for projects with partial job progress reporting)
Sample Message Log Entries:
15/12/2007 10:09:37|CPDN|Sending scheduler request: To send trickle-up message. Requesting 0 seconds of work, reporting 0 completed tasks
The speed of reporting is impacted e.g. by the Connection Time set in the Device profile and other client scheduling activities. When a Connect Time of 2 days is set in order to be able to buffer more work, there is no rush to update the server side scheduler, thus the RtR clearing would be postponed up to when the 'Connect Time' of 2 days has passed. When a permanent internet connection does exist though and is open, the client will try to push out the Task Result Files, so at least a backup exists on the server side, still there being no need to update the server scheduler. The client did set that Connect Time after all.
There is of course the Projects Tab 'Update' button that opens / tries opening a connection for 5 minutes with the server to push out any results and force the RtR's. This usually results in the desired action. Better is to opt for the WCG recommended short connect time defaulting to 0.3 days (7.2 hours) and use the 5.10.x release local 'Additional Buffer' days local preferences option set obtain extra work in the Tasks queue. This will clear the RtR's more frequently without unnecessary extra Scheduling Daemon contacting.
When RtR's are not cleared or new work is not fetched, it is possible that the client got confused. Frequent changing of profile settings & local preferences and repeated panic states (see discussions on Earliest Deadline First / High Priority processing) can cause such conditions to develop. BOINC needs 1 or 2 weeks to find it's balance and understand *particularly* if multiple projects are attached to the client. Best is to decide on what's wanted, make the changes and leave it be. Resist the urge to micromanage
*This info credited to Sekerob, Community Advisor @ WCG Forums