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

Quick Go ยป
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 1500 times and has 3 replies Next Thread
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
UGM1: One 'No Reply' initiating 2 extra repair copies

A wasteful setup, regardless if a result is using 50KB or 1500KB storage. At 1/5/15 23:17:20 the _0 copy goes NR and 2 more copies are distributed, one gets 'Server Aborted', by fortune, because the second extra copy had not been started yet and the first happened to connect to the server for whatever reason to receive the instruction. Would the task have started before hand, there'd been 3 valid computed copies.

Project Name: Uncovering Genome Mysteries
Created: 12/28/2014 14:52:12
Name: ugm1_ugm1_05417_1835
Minimum Quorum: 2
Replication: 2


Result Name App Version Number Status Sent Time Time Due /
Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit
ugm1_ ugm1_ 05417_ 1835_ 2-- 723 Valid 1/5/15 23:18:27 1/6/15 11:18:36 3.62 83.4 / 101.3
ugm1_ ugm1_ 05417_ 1835_ 3-- 723 Server Aborted 1/5/15 23:18:26 1/6/15 11:43:18 0.00 0.0 / 0.0
ugm1_ ugm1_ 05417_ 1835_ 1-- 723 Valid 12/29/14 23:17:20 1/6/15 07:30:02 4.94 119.2 / 101.3
ugm1_ ugm1_ 05417_ 1835_ 0-- - No Reply 12/29/14 23:17:20 1/5/15 23:17:20 0.00 0.0 / 0.0
----------------------------------------
[Edit 2 times, last edit by Former Member at Jan 6, 2015 12:38:29 PM]
[Jan 6, 2015 12:35:26 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: UGM1: One 'No Reply' initiating 2 extra repair copies

_0 and _1 were sent out at the same time, so both will have gone No Reply at the same time, hence 2 repair copies. Then _1 completed successfully, albeit late. It's a compromise choice between grid efficiency and progressing the research quickly. It would be interesting to know what proportion of such cases end up with a superfluous copy being validated, but no, Keith U, don't divert your effort from alpha testing OET :)
[Jan 6, 2015 2:01:37 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: UGM1: One 'No Reply' initiating 2 extra repair copies

Had not noticed both were overdue, _0/_1. The logic could include if there was intermediate contact with the late hosts, even if the late task was already running, and it should have been running in high priority on top of that.

Yes, WCG is very good at piling up many small items to improve and fix to make the experience livelier for the happy feet ;>)
[Jan 6, 2015 3:01:36 PM]   Link   Report threatening or abusive post: please login first  Go to top 
uplinger
Former World Community Grid Tech
Joined: May 23, 2005
Post Count: 3952
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: UGM1: One 'No Reply' initiating 2 extra repair copies

Sekerob,

What I think may have happened is this.

Two original results were sent out. Both came back as no reply by the deadline and it triggered another 2 results to be sent. One of the first two came back shortly after deadline and was waiting for pending validation. There is a chance that 3 valid results could be returned in this case, but it is such a small fringe case.

The result is not in the database anymore so I can't check the return times.

Thanks,
-Uplinger
[Jan 9, 2015 5:00:09 PM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread