Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 4
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 1538 times and has 3 replies Next Thread
NixChix
Veteran Cruncher
United States
Joined: Apr 29, 2007
Post Count: 1187
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
confused Problem with Back-Off Timer

I am encountering a problem when I try to fill my queue with jobs from a project with a steady flow and switch back to HSTB or FAH2.

After performing an manual update I see "Deferred for: 00:01:20" under Status in BoincTasks. After watching it count down to 00:00:00 and then disappearing, I perform another manual update and instead I am rebuffed with the message from the server "Not sending work - last request too recent: 89 sec".

Another thing I have noticed that one of my 3 computers starts counting down at 00:02:00, the other 2 start at about 00:01:20 (40 seconds less). I do not have a problem with the computer that starts counting from 2 minutes, only the 2 that start at 1:20. It would seem that either the server sends incorrect backoff value, they receive or report to BoincTasks in correctly, or BoincTasks is getting it wrong.

I performed a manual update at the 2 computers using BOINC Manager and they start counting down at 2 minutes. It didn't think it could be the server and that proves it. I compared the message on BOINC Manager at the computer to that reported by BoincTasks and there is a discrepancy. The only potential spots are transmitting the status to BoincTasks by the client computer or BoincTasks itself.

I have Boinc 7.2.47 (current WCG recommendation) installed on all three computers and the the latest version of BoincTasks (1.69) installed. This is not a new development. I have noticed it for a while, but finally am annoyed enough to look into it.

Any thoughts? thinking

Cheers coffee
----------------------------------------

[Apr 15, 2017 7:23:50 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Problem with Back-Off Timer

Learn to accept there's a messaging queue with the scheduling daemons. The standard WCG backoff is 121 seconds, but the time between initializing a request and completing and the back-off tracker being told to let the next one thru is fluid. What you see in BOINCTasks of course feeds off the BOINC client, and that is not necessarily real-time... that app has a messaging queue too, RPC... the more clients attached, the more latency is introduced.
[Apr 16, 2017 10:16:55 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Problem with Back-Off Timer

PS, Rosetta's is 242 seconds, double WCG... a coincidence maybe or some multiplier in the system the project admin can knob.
[Apr 16, 2017 10:19:33 AM]   Link   Report threatening or abusive post: please login first  Go to top 
NixChix
Veteran Cruncher
United States
Joined: Apr 29, 2007
Post Count: 1187
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Problem with Back-Off Timer

Strange, this problem spontaneously corrected itself before I was able to even try to diagnose. It has been wrong for a long time and now it is not there. thinking

Cheers coffee
----------------------------------------

[Apr 18, 2017 7:12:32 AM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread