| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Maximum Number of network connection |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 5
|
|
| Author |
|
|
Wilgard
Advanced Cruncher Joined: Dec 15, 2005 Post Count: 83 Status: Offline Project Badges: 
|
Hello,
----------------------------------------I am going to hollidays during 9 days and I would like my computer work during this time. But sometimes I have some problems of Internet connection and I want to configure my Boinc Agent (wcg) for working during 9 days as if I will not have an Internet connection. So, I have configure my wcg options with a "Connect to network about every" at "10.0" days. But I do see more project (only 2 : 1 in calculate state and 1 in waiting state). So, my 2 questions are: - "what is the maximum value of : Connect to network about every ? " (1.0, 2.0, .... , 10.0) - "HOw can I do for configure my Boinc Agent for workning during 9 days without connection ?" Thank you for your help . Wilgard  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The short answer is, you can't.
Work units time out in a week, so you should never be assigned more work units than you can complete in seven days. Once you have changed the setting to a reasonable value (I suggest 5 days) then it will take a little while for the settings to filter through and for BOINC to request more work. You can speed this up by clicking the "update" button in BOINC Manager, but it still won't be immediate. Perhaps you can find a utility to monitor your network connection and refresh it periodically? Of course, this won't help if your router needs a cold restart, but even this can often be automated. For example, I have a WebSTAR cable modem, and there is a page where you can restart it: http://192.168.100.1/gscan.htm - you could set up a script to send the required GET request directly to your modem without too much trouble. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
hmmm now i'm wondering....192.168.100.1 shows up as my default gateway in my d-link wireless router....if i blank that field and save, it just comes back. It wont let me do anything to it, yet it pings with about 241ms average, so it seems to go way outside my ISP. Putting anything else in the field, on save blanks the field....it only accepts 192.168.100.1. Searching the router config pages, there is a script: javascript:commandset('restart'), which is probably same as the gscan.htm. Run that once in a while if comms is poor.
----------------------------------------On disconnecting, looking thru the BOINC logs, last night there was a commsinterruption with WCG, while it was trying to upload results. It re-attempted uninterruptedly every minute from 4:21am CET to 5:46am CET. I cant find the post here relating to this endless retry, but think there was a reason why its not backing off for an ever increasing period. That's one feature that should be build into the Agent. a 06-07-30 5:43:50|World Community Grid|Temporarily failed upload of B00139_0282_CTMA3A2-10-3-22_2_2: http error b 06-07-30 5:43:50|World Community Grid|Backing off 1 minutes and 0 seconds on upload of file B00139_0282_CTMA3A2-10-3-22_2_2 c 06-07-30 5:43:50|World Community Grid|Temporarily failed upload of B00139_0282_CTMA3A2-10-3-22_2_3: http error d 06-07-30 5:43:50|World Community Grid|Backing off 1 minutes and 0 seconds on upload of file B00139_0282_CTMA3A2-10-3-22_2_3 e 06-07-30 5:45:29|World Community Grid|Task B00139_0280_CTMA3A2-10-3-2_0 exited with zero status but no 'finished' file f 06-07-30 5:45:29|World Community Grid|If this happens repeatedly you may need to reset the project. g 06-07-30 5:45:40|World Community Grid|Started upload of file B00139_0282_CTMA3A2-10-3-22_2_0 h 06-07-30 5:45:51|World Community Grid|Started upload of file B00139_0282_CTMA3A2-10-3-22_2_1 i 06-07-30 5:46:32|World Community Grid|Restarting task B00139_0280_CTMA3A2-10-3-2_0 using hdc version 505 j 06-07-30 5:46:40|World Community Grid|Finished upload of file B00139_0282_CTMA3A2-10-3-22_2_0 k 06-07-30 5:46:40|World Community Grid|Throughput 11004 bytes/sec l 06-07-30 5:46:40|World Community Grid|Finished upload of file m B00139_0282_CTMA3A2-10-3-22_2_1 n 06-07-30 5:46:40|World Community Grid|Throughput 4641 bytes/sec o 06-07-30 5:46:40|World Community Grid|Started upload of file B00139_0282_CTMA3A2-10-3-22_2_2 p 06-07-30 5:46:40|World Community Grid|Started upload of file B00139_0282_CTMA3A2-10-3-22_2_3 q 06-07-30 5:46:45|World Community Grid|Finished upload of file B00139_0282_CTMA3A2-10-3-22_2_2 r 06-07-30 5:46:45|World Community Grid|Throughput 2057 bytes/sec s 06-07-30 5:46:45|World Community Grid|Finished upload of file B00139_0282_CTMA3A2-10-3-22_2_3 t 06-07-30 5:46:45|World Community Grid|Throughput 3712 bytes/sec Anyway, when going away for trips <= 7 days, i take the BOINC suspend network option....sure don't want that box to connect while absent and take my livelihood with it due some hacker/zombie or other trash roaming the Internet.....why 7 days. Well after the 7th day a 4th copy is send out for crunching. Thus when one comes back and reconnects after 9, u'd effectively have hi probability of returning duplicated work. Albeit, just on the border on the 8th day, for force majeure reasons, one would still get the credit of the quorum, if that had already been established. As for the script modem reset.....no i'm not going to leave the wifi router broadcasting for 7 days and some punk burgling.....the 'how to' been published on the web. Come to think....if u want to keep your PC happy while traveling longer than 7, temp install UD. That agent allows a 3 week return...it still would fall in pause after a few days.......but for that there is the UD_Monitor answer.....that can cache as many slots/wu's as u like.
WCG
Please help to make the Forums an enjoyable experience for All! |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Sekerob, that address isn't routable - it isn't going to leave your network.
The backoff was removed from the agent for the case when it's a local network failure so that the agent can get back to work as soon as the connection is back. The only reason for the backoff is to avoid overloading the server if the server has been down and there is a large backlog of scheduler requests. Since the server is just fine, there's no reason to backoff. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Didactylos, I'm baffled over this concern of not being able to crunch for a minute quote "...so that the agent can get back to work as soon as the connection is back." unquote. Why is it acting as if oblivious of the fact that cached units might be there beyond the one its crunching currently and oblivious of a profile setting of scheduled contacts?
----------------------------------------Dark waters run deep.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|