@ |
Home | Help | Status | Forums | Glossary | Account
|
log in |
Message boards : Help! : Uploads stuck
Author | Message |
---|---|
Me G5 Mac has four completed yello tasks that ’ave been stuck at “Uploadin’” fer a few days. They be marked “No reply” in th’ database: should I abort them at me end? Would ye like me t’ try ’n copy th’ results afore doin’ so? | |
ID: 8719 | Rating: 0 | rate: / Reply Quote | |
Go ahead and abort them. There was not work for Mac in this round, though some machines got told otherwise. The older app version is in the database, but the | |
ID: 8720 | Rating: 0 | rate: / Reply Quote | |
Also, I think the IP address changed. I had to reboot my windows machines in order for them to be able to contact the project. | |
ID: 8721 | Rating: 0 | rate: / Reply Quote | |
zombie67 wrote: Also, I think the IP address changed. I had to reboot my windows machines in order for them to be able to contact the project. Yes, it did. That could explain it. One morning, after having the same IP address for ages, it changed. But you should not have to reboot to deal with that once the change is in DNS (which it was a few hours later). ____________ -- Eric Myers "Education is not the filling of a pail, but the lighting of a fire." -- William Butler Yeats | |
ID: 8722 | Rating: 0 | rate: / Reply Quote | |
Don't know what to say, but it did not work like that. They could not find the new address for several days. The only thing that worked was rebooting. This is with xp64, and 6.10.6. | |
ID: 8723 | Rating: 0 | rate: / Reply Quote | |
I wonder if we've demonstrated another bug in BOINC? | |
ID: 8724 | Rating: 0 | rate: / Reply Quote | |
Wormholio wrote: Go ahead and abort them. Aye, aye. There was not work for Mac in this round, though some machines got told otherwise. The older app version is in the database, but the executable was not in the download area. If you had it already, fine, but if not, it would fail to download. They had definitely run, I guess from an app I’d previously looted. 'n a few tasks were uploaded (& reported) earlier in th’ run. ____________ Was that a Siren I heard? | |
ID: 8725 | Rating: 0 | rate: / Reply Quote | |
Odysseus wrote:
In that case it seems reasonable to assume that the IP address change of the project explains the upload problems. ____________ -- Eric Myers "Education is not the filling of a pail, but the lighting of a fire." -- William Butler Yeats | |
ID: 8726 | Rating: 0 | rate: / Reply Quote | |
Wormholio wrote: I wonder if we've demonstrated another bug in BOINC? No. It happened on SETI also. They believed it was some library bug if I remember correctly, and not something that could be easily fixed with programming. | |
ID: 8727 | Rating: 0 | rate: / Reply Quote | |
I’ve aborted th’ tasks, but I still can nah seem t’ get them reported: Mon 28 Sep 18:14:47 2009|Pirates@Home|Fetching scheduler list Mon 28 Sep 18:16:03 2009||Project communication failed: attempting access to reference site Mon 28 Sep 18:16:04 2009||Access to reference site succeeded - project servers may be temporarily down. But accordin’ t’ th’ status page, th’ servers are up. Be thar no way t’ get BOINC t’ use a new IP without rebootin’ th’ system? ____________ Was that a Siren I heard? | |
ID: 8728 | Rating: 0 | rate: / Reply Quote | |
Odysseus wrote:
Once I realized the IP address had changed I edited DNS to propoagate the change. It can take up to an hour for those changes to propagate. But then it is available to your computer. Is it not connecting to the project? If BOINC saves the IP address instead of checking again after the TTL has expired (TTL stands for Time To Live) then it would continue to try the old IP address. If your client won't connect to the project, you could try stopping it (the client core, not just the manager) and restarting, to see if that clears it, instead of having to reboot the machine (which seems extereme). ____________ -- Eric Myers "Education is not the filling of a pail, but the lighting of a fire." -- William Butler Yeats | |
ID: 8729 | Rating: 0 | rate: / Reply Quote | |
Wormholio wrote:
No. ’twas gettin’ 24-hour deferrals on each failed connection attempt, today Tue 29 Sep 18:16:08 2009|Pirates@Home|Fetching scheduler list Tue 29 Sep 18:17:24 2009||Project communication failed: attempting access to reference site Tue 29 Sep 18:17:25 2009||Access to reference site succeeded - project servers may be temporarily down. If your client won't connect to the project, you could try stopping it (the client core, not just the manager) and restarting, to see if that clears it, instead of having to reboot the machine (which seems extereme). Aye, that's all it took. Relaunchin’ BOINC (I ’ave th’ client installed as an application, nah a service, on this machine, so quittin’ Manager stops th’ CC processes as well): Tue 29 Sep 18:31:53 2009|Pirates@Home|URL: http://pirates.spy-hill.net/; Computer ID: 24222; location: school; project prefs: default I wasn’t about t’ wait 23+ hours, so I hit Update: Tue 29 Sep 18:33:44 2009|Pirates@Home|Fetching scheduler list Tue 29 Sep 18:33:49 2009|Pirates@Home|Master file download succeeded Tue 29 Sep 18:33:54 2009|Pirates@Home|Sending scheduler request: Requested by user. Requesting 3718 seconds of work, reporting 4 completed tasks Tue 29 Sep 18:33:59 2009|Pirates@Home|Scheduler request succeeded: got 0 new tasks Tue 29 Sep 18:33:59 2009|Pirates@Home|Message from server: Starting in 183 days and 3 hours, project will require a minimum BOINC core client version of 6.1.0. You are currently using version 5.10.45; please upgrade before this time. [color=red]Tue 29 Sep 18:33:59 2009|Pirates@Home|Message from server: (Project has no jobs available)[/color] I suppose I could ’ave done that instead o’ abortin’ th’ tasks, ’n they’d ’ave been uploaded. Oh, well: I doubt they’ll be sorely missed! P.S. Didn’t th’ format invoked by th’ code tag use t’ scroll horizontally? Had t’ unstretch th’ thread here … ____________ Was that a Siren I heard? | |
ID: 8730 | Rating: 0 | rate: / Reply Quote | |
So I think this indicates a bug in BOINC. At the time of the upload, it should look up the IP address in DNS and use that for the connection, instead of using some previous, possibly stale value. | |
ID: 8731 | Rating: 0 | rate: / Reply Quote | |
BOINC 6.x may handle thin’s differently. I’ll get around t’ updatin’ eventually … | |
ID: 8732 | Rating: 0 | rate: / Reply Quote | |
This bug has been around for a LONG time. Every so often a project will start up using a dynamic DNS, and this forces them to change to a static address. It was supposed to have been fixed a while back. A library changed for something. But it didn't solve the problem. | |
ID: 8733 | Rating: 0 | rate: / Reply Quote | |
zombie67 wrote: This bug has been around for a LONG time. Every so often a project will start up using a dynamic DNS, and this forces them to change to a static address. It was supposed to have been fixed a while back. A library changed for something. But it didn't solve the problem. cURL? ____________ Kathryn :o) The BOINC FAQ Service The Unofficial BOINC Wiki The Trac System | |
ID: 8734 | Rating: 0 | rate: / Reply Quote | |
That sounds familiar. Probably the thing. | |
ID: 8735 | Rating: 0 | rate: / Reply Quote | |
Message boards : Help! : Uploads stuck
Home | Help | Status | Forums | Glossary | Account
|