Feeder not running
log in

Advanced search

Message boards : Number crunching : Feeder not running

Previous · 1 · 2
Author Message
Darrell
Avatar
Send message
Joined: 30 May 17
Posts: 10
Credit: 4,200
RAC: 0
Message 389 - Posted: 6 Sep 2017, 16:37:24 UTC

Allowed new tasks, but will be a couple of hours before Boinc is ready to request a task.
____________

Jim1348
Send message
Joined: 4 May 17
Posts: 32
Credit: 515,350
RAC: 121
Message 390 - Posted: 6 Sep 2017, 18:14:09 UTC - in response to Message 389.

You won't get anything until he gets the domain problem fixed.

Profile PDW
Send message
Joined: 17 Apr 17
Posts: 4
Credit: 206,050
RAC: 139
Message 391 - Posted: 6 Sep 2017, 18:21:52 UTC - in response to Message 390.

Not true, I already have some :-)

Jim1348
Send message
Joined: 4 May 17
Posts: 32
Credit: 515,350
RAC: 121
Message 392 - Posted: 6 Sep 2017, 19:15:46 UTC - in response to Message 391.

I suspended all other tasks, and haven't gotten any. You must have hit them at just the right time. But I do see several in progress on the server status page, but with zero users. That is quite a trick in itself.

Werinbert
Send message
Joined: 25 Jul 17
Posts: 2
Credit: 150
RAC: 0
Message 393 - Posted: 6 Sep 2017, 19:17:21 UTC - in response to Message 391.

Not true, I already have some :-)

So what is different in your set up that allows you to get tasks where the rest of us are getting server down messages? I have tried resetting the project and all that.

Profile PDW
Send message
Joined: 17 Apr 17
Posts: 4
Credit: 206,050
RAC: 139
Message 394 - Posted: 6 Sep 2017, 19:24:41 UTC - in response to Message 392.

I suspended all other tasks, and haven't gotten any. You must have hit them at just the right time. But I do see several in progress on the server status page, but with zero users. That is quite a trick in itself.

I think the user count only counts users that have returned results, so though some users have downloaded tasks to process no-one has returned any.

Profile PDW
Send message
Joined: 17 Apr 17
Posts: 4
Credit: 206,050
RAC: 139
Message 395 - Posted: 6 Sep 2017, 19:27:16 UTC - in response to Message 393.

Not true, I already have some :-)

So what is different in your set up that allows you to get tasks where the rest of us are getting server down messages? I have tried resetting the project and all that.

Nothing as far as I know.
I upgraded my VBox to 5.1.26, then added the project climateathome.info/climateathome/ to the client and down they came :-)

The one I have running isn't working though, well not doing anything worthwhile at the moment. Much like how they would be when it last ran :-(

Jim1348
Send message
Joined: 4 May 17
Posts: 32
Credit: 515,350
RAC: 121
Message 396 - Posted: 6 Sep 2017, 20:04:24 UTC - in response to Message 392.
Last modified: 6 Sep 2017, 20:09:29 UTC

I suspended all other tasks, and haven't gotten any. You must have hit them at just the right time.

Well guess what? While I couldn't get work on my Ubuntu 16.04 machine, which I normally use for this project, it downloaded one work unit right away on my Windows 7 64-bit machine. Both are using VirtualBox 5.1.26. So I guess there is a difference in VBox after all.

That is true on one other project I know of, USPEX, though there they specifically tell you that it requires Windows. So there is something new every day in this business.

EDIT: It is running the new Version 25.00, which seems OK, but I had no problems with the old version 24.00 unless I rebooted. I have not tried that yet, and would like to complete one before I do.

Kai Liu
Project administrator
Project developer
Project scientist
Send message
Joined: 29 Mar 17
Posts: 101
Credit: 122,103
RAC: 0
Message 397 - Posted: 6 Sep 2017, 20:13:12 UTC - in response to Message 396.

Thanks. I have fixed the daemon thing with the help of Nicolás Alvarez from BOINC. The BOINC has a html/user/schedulers.txt which saved URL for scheduler. In my previous version, I used IP instead of domain in that file.

Currently version 25.0 only supports windows and have a few tasks. Let's wait and see and then I will adjust credits, create more tasks, then add Linux version.

Jim1348
Send message
Joined: 4 May 17
Posts: 32
Credit: 515,350
RAC: 121
Message 398 - Posted: 6 Sep 2017, 21:28:03 UTC - in response to Message 397.

The CPU percentage started out slowly, as is usual for VBox projects, and increased to maybe 20% or so after 20 minutes. But then it started to fall, and is down below 5% after 1 1/2 hours. However, the estimated time remaining is still decreasing at the expected rate, so I can't say it is dying yet. I will just let it run and see if it finishes, with the estimated time remaining being 1 day 18 1/2 hours from now.

Darrell
Avatar
Send message
Joined: 30 May 17
Posts: 10
Credit: 4,200
RAC: 0
Message 399 - Posted: 6 Sep 2017, 22:12:30 UTC

Got a task and it has been running for two hours now. Percent complete is 2.6 and slowing climbing, estimated time to completion is 2 days, 7 hours. I will give it a few more hours and then shut the Boinc client down and see if the task restarts from where it left off. The virtual box machine was created in the directory given to the task to run in so that problem is solved. If the task restarts correctly from a checkpoint, I think you'll have a winner.
____________

Profile Bill F
Send message
Joined: 5 May 17
Posts: 14
Credit: 450
RAC: 0
Message 400 - Posted: 7 Sep 2017, 1:30:04 UTC - in response to Message 397.

Kai

When you get a moment you might want to post something on "News" regarding your relocation the changed IP and the fact that a Version 25 has been released to a limited group of OS's

Thanks
Bill F

P.S. My current Ver. 25 WU is at 2% and is moving along nicely so far.
____________

Jim1348
Send message
Joined: 4 May 17
Posts: 32
Credit: 515,350
RAC: 121
Message 401 - Posted: 7 Sep 2017, 2:01:26 UTC - in response to Message 400.

My version 25 work unit is also at 2% after six hours, but not exactly racing. The remaining time left is reducing slower than real time, and I hope it makes it to the end. But hopefully the efficiency can be increased; many VBox projects I work on are over 90%.

Darrell
Avatar
Send message
Joined: 30 May 17
Posts: 10
Credit: 4,200
RAC: 0
Message 402 - Posted: 7 Sep 2017, 3:21:06 UTC

Shutdown the client. Restarted the computer. After Boinc was restarted, the task has failed to restart and is postponed due to VM Hypervisor failing to restart in a timely manner.

Link to stderr.txt => https://1drv.ms/t/s!ArIvftV8roEagXcrQDwoLX6ru-Mc

Screenshot of VM Box manager => https://1drv.ms/i/s!ArIvftV8roEagXijjnF1rJYF3FTK

Will see what happens when Boinc tries to restart it again.
____________

Darrell
Avatar
Send message
Joined: 30 May 17
Posts: 10
Credit: 4,200
RAC: 0
Message 403 - Posted: 7 Sep 2017, 21:00:30 UTC

Tried suspending other project's running tasks but to no avail. Could not get Boinc to attempt to restart the postponed task. So shut down Boinc and restarted the computer. After the restart and starting Boinc, the task attempted to restart and errored out. Here is a link to the task:

https://climateathome.info/climateathome/result.php?resultid=16237

Hopefully, from its output, you can determine why it wouldn't restart.
____________

Jim1348
Send message
Joined: 4 May 17
Posts: 32
Credit: 515,350
RAC: 121
Message 404 - Posted: 9 Sep 2017, 22:07:01 UTC - in response to Message 401.

My work unit completed after 3 days 9 hours on a lightly-loaded i7-4771. I avoided all interruptions and did not suspend it at any time. The CPU usage was only 1.2% at the end.

Kai Liu
Project administrator
Project developer
Project scientist
Send message
Joined: 29 Mar 17
Posts: 101
Credit: 122,103
RAC: 0
Message 405 - Posted: 9 Sep 2017, 22:09:42 UTC - in response to Message 404.

My work unit completed after 3 days 9 hours on a lightly-loaded i7-4771. I avoided all interruptions and did not suspend it at any time. The CPU usage was only 1.2% at the end.


Thanks for your message. I will check the results later.

Previous · 1 · 2

Message boards : Number crunching : Feeder not running


Main page · Your account · Message boards


Copyright © 2017 NSF I/UCRC for Spatiotemporal Thinking, Computing, and Applications, George Mason University, NASA