| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: new response to 'update request' |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 5
|
|
| Author |
|
|
grayon
Cruncher Joined: Nov 16, 2004 Post Count: 18 Status: Offline Project Badges: 
|
I now have 4pgs of "detached" result status due to 'task aborted by server' response to an update request. Was it something I said !!!??? &^%*($#
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
No, just the color of your eyes ;>)
Seriously... scan the stdoutdae.txt file in the BOINC data directory for the sequence of events. Normally "tasks aborted by sever" would lead to them being listed on the Result Status page as "server aborted" or (falsely) "error" with a client v7, but in this case it hints at your host having actually been detached from WCG, which then in the second instance leads to the server confirming back that the tasks were aborted due to a detach event. Client attached to BOINCstats Account Manager or GridRep? Those can lead to conflicts with WCG. |
||
|
|
grayon
Cruncher Joined: Nov 16, 2004 Post Count: 18 Status: Offline Project Badges:
|
Sorry; I should have given more details....
configuration: 4 e3-1200 servers....Fedora 18...KDE... all localhost.localdomain problem occurs when I attempt an update with something like 8 to 16 tasks ready to report. wcg's server handles the 1st system fine, but if i switch to the next system, and it too has 8 to 14 it flushes all my 'ready to report' plus all 'running' status tasks down the tubes.. it appears that the wcg's server can't handle back to back to back multitask updates from different clients with the same name and the same source. I know having the same client name may complicate things but it should be handled!!! come-on guys I'm loosing alot tasks, hours...not to mention the almighty points. If it can be done, the server should handle it !!!!!!!!! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
"wcg's server handles the 1st system fine, but if i switch to the next system", suggests the servers are getting confused by what you do or how you have configured the 2 systems. Hopefully the 2 are not clones with same System network ID, and through this 'switching' are not using the same LAN IP. In that case the servers in a nutshell conclude: Hold on, that work was not assigned to that device, abort and assign fresh work. First thing, ensure the hostnames are unique and do refresh IPs, e.g. pre-assign / fix a different IP for each device [My Router has a table locked to the different MAC codes, so DHCP always assigns a unique/same IP to each host on the LAN.
|
||
|
|
grayon
Cruncher Joined: Nov 16, 2004 Post Count: 18 Status: Offline Project Badges:
|
temporary fix...cache=0.0
|
||
|
|
|