| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Work may be reduced for long periods when server is down |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 44
|
|
| Author |
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
Jean probably means the global_prefs_override.xml. Yes you can edit latter directly and use the advanced menu to read those settings from version 5.8. Even simpler than that!  Since I-don't-remember-when the number of CPUs is offered for update in the "Advanced-->Preferences" boxes. Was true at least for 5.10.42 and now 5.10.45. Cheers. Jean. ---------------------------------------- [Edit 1 times, last edit by JmBoullier at Mar 12, 2008 6:31:50 PM] |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
Since "my" thread "Problem after msg "Server can't open database" has been locked as duplicate of this one I put the above link for easier reference from here.
----------------------------------------Kremmen, the name of the file that knreed would have liked to see is at the beginning of my post, unfortunately it contains only the latest reply from the server. |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
It has happened again this evening...
----------------------------------------But since I was not there Boinc could re-issue the fetch by itself just one hour later. It looks like I discovered the problem 5 minutes too soon this afternoon. So it seems that the problem is less definitive than I thought. For the time being I will simply add a little more extra work to my preferences to give it enough time to recover without falling short of WUs if/when it happens again. However, perhaps one hour is a little too long for the retry? Cheers. Jean. PS: By the way I have asked Tedi to move this thread to the Boinc Support forum. |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
I just finished tracking down the bug. Unfortunately it will have to be fixed by a client update so it will be something that will probably not be fixed until the 6.0 client.
----------------------------------------Here is the BOINC track ticket for the bug: http://boinc.berkeley.edu/trac/ticket/578 In the meantime, I will be updating the code on the server in the morning (US time) so that in the cases that cause this problem, instead of sending a message to the client to try again in an hour, it will instead send the backoff delay so that it will try again in 1 minute and then slowly increase the backoff time from there. This should reduce the time between experiencing this error and getting the correct venue re-established. What is happening to your computers, is that when this error is given, instead of your computer using the preferences for the venue that you have assigned to the computer, it will instead use your 'default' preferences. [Edit 1 times, last edit by knreed at Mar 13, 2008 1:33:38 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Two things strike me about this:
1) My twin-processor machine has been working for months on WCG and the only time (that I know of) that this has happened was twice in the last couple of days. That seems to suggest that something has suddenly become unreliable in the WCG scheduler. 2) Wouldn't not responding at all if the database is inaccessible be better than responding with something incorrect? The client gets no useful information anyhow if the database is down. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The Server software was upgraded about 2 weeks ago to 6.01
----------------------------------------Try setting the CPU cores, as per the discussion earlier in this thread, in the global_prefs_override.xml (... thru 5.8) or using the Advanced menu > Preferences screen (from 5.10). I've set it locally since there is a P4HT which needs to stick to 1 core. Looked thru the stdoutdae.txt logs and cant see this happening here on any host. The thing is that each time the BOINC client detects different core permissions it will do a benchmark test and not one every 5 days.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 3 times, last edit by Sekerob at Mar 13, 2008 6:19:26 AM] |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Two things strike me about this: 1) My twin-processor machine has been working for months on WCG and the only time (that I know of) that this has happened was twice in the last couple of days. That seems to suggest that something has suddenly become unreliable in the WCG scheduler. 2) Wouldn't not responding at all if the database is inaccessible be better than responding with something incorrect? The client gets no useful information anyhow if the database is down. The answer to both of these is related. Previously if the scheduler could not open a database connection , then scheduler basically crashed. This had the effect of not sending a response back to the user as you have suggested in #2. |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Ok - found a way to fix this. I have updated the server so that the tag <project_is_down/> is sent by the scheduler if connection to the database cannot be established. The client will then not take any action based on the reply and therefore not change the venue for the computer.
Let me know if anyone experiences any further problems. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ok, so ALL my clients have this message:
3/14/2008 4:41:24 PM|World Community Grid|Reporting 2 tasks 3/14/2008 4:41:30 PM|World Community Grid|Scheduler request succeeded 3/14/2008 4:41:30 PM|World Community Grid|Message from server: Server can't open database 3/14/2008 4:41:30 PM|World Community Grid|Project is down 3/14/2008 4:41:56 PM|World Community Grid|Reporting 1 tasks 3/14/2008 4:42:02 PM|World Community Grid|Scheduler request succeeded 3/14/2008 4:42:02 PM|World Community Grid|Message from server: Server can't open database 3/14/2008 4:42:02 PM|World Community Grid|Project is down ...and nothing is reported. |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Yes - there was a period of 8 minutes today when some clients would have gotten that message.
Can you please post a follow up and let us know if at around: 3/14/2008 4:48:02 PM they all connected just fine? |
||
|
|
|