Questions and Answers :
Unix/Linux :
Long waiting time after connection failure
Message board moderation
Author | Message |
---|---|
wobo Send message Joined: 3 Apr 99 Posts: 25 Credit: 1,957,899 RAC: 0 |
Last tuesday I went to Paris, France, for 4 days and thought, my boinc may have a nice undisturbed crunching time. The machine runs on Linux and is used to long uptimes. I have a permanent DSL connection, so everything was set. When I returned I realized that 2 hours after I left, the internet connection was cut and while trying to reconnect my router died. So BOINC was trying to connect for 4 days and of course it added up the waiting time between new connection attempts. Now it's on a waiting loop of 2 days and some hours and I can't cut that off, although the internet connection is on again. Whatever I do (Ctrl-C & restart, update of the version (4.19), nothing stopps BOINC to count down the waiting loop: "Deferring communication with project for 2 days, 10 hours, 27 minutes, and 26 seconds". Is there a solution to this? IMHO it's a bug. One of the core features of BOINC is to run during idle times. What's more 'idle' than night time or weekend? But it can always happen that during this idle time the router dies or the cat eats the data line. So there must be a way to stop this waiting loop. If not, it's a bug. wobo |
Trane Francks Send message Joined: 18 Jun 99 Posts: 221 Credit: 122,319 RAC: 0 |
Two possible fixes: 1) Restart the client with --update_prefs 2) close BOINC, edit client_state.xml to change the RPC time to 0, then restart The waiting period is a design feature that attempts to "trickle in" connections after an outage so as not to overwhelm and bring down a server on large projects. |
wobo Send message Joined: 3 Apr 99 Posts: 25 Credit: 1,957,899 RAC: 0 |
> 1) Restart the client with --update_prefs Did that, no difference > 2) close BOINC, edit client_state.xml to change the RPC time to 0, then > restart Ah, that should have done it. Thanks for the info, I'll use it next time. Meanwhile I followed an advice I received from a member of the Mandrakeuser.de team: I did a --detach and after that a --attach. That brought me out of the loop but also created a new client. Hmm, of course that's no problem because both clients add up and only the new one is in use. Locally I still have only one client, one directory. And in my stats I see that the 2 unfinished WUs are still in "being processed" status, although they don't show up anywhere on my harddisk. I guess those 2 are lost here. wobo |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.