| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Beta Testing
Forum: Beta Test Support Forum
Thread: New Set of Beta WU's - FAAH |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 484
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Are these betas being sent only to reliable hosts, or to any hosts signed up for beta testing? If it's only reliable hosts, then another nice thing is that the errors aren't knocking the hosts out of reliable status. I'm continuing to get new betas on the machines that suffered -161 errors. :) Beta's go to all machines, regardless reliability state... to also test those not so reliable machines, to see if they're up to the [new project] jobs. All it needs is to opt in. :D Think the production reliability is not affected by Beta fails, at least through this had true Beta errors and still getting production short-deadline tasks. --//-- edit: Seeing in BOINCTasks history three -161 now and 2 server aborted Betas. [Edit 2 times, last edit by Former Member at Jul 2, 2011 4:48:56 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Think the 6.10.59 scheduler is a little lost, suffering extensive dichotomy... now 15 tasks have run partial, 7 of which are CEP2, 4 of these progressing in High Priority. Linux is holding up, but I've never seen BOINC using 2.9GB of Virtual memory, total Disk use including temp files at 5.13GB on top of 2.88GB of ram (info per BOINCTasks and 6.12.33 remotely connected BOINC Manager. Hope the seams hold.
--//-- |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: 
|
I would stay away from that system for a while; poking the KB never helps when your drive is being used for memory. Alternatively suspending non-started tasks would prevent the situation further deteriorating, and when tasks finish things will get easier on the system. Pity Boinc can't see a disaster coming.
At least there is no chance of your 15 partial runs being server aborted ;) Are we to take it that this Beta run is at an end? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Seen only 4 of 6 active sciences. That said, Clean Water is not really of ''loading'' interest because the only thing that's coming down is an instruction to compute x/y/z seed from the table already on the hosts. HCMD2 is also not that hefty, but the exercise would have to be repeated in full come Tuesday cause ''the load could not be simulated'' ;P
As for suspending, done that this morning to force those CEP2 tasks ahead that had 5-6-7 hours down, to get the quickest stress relieve, but from there I'll just let it... think the machine will survive without too much trouble. Set up uses 2 physical disks to spread load. --//-- |
||
|
|
johncmacalister2010@gmail.com
Veteran Cruncher Canada Joined: Nov 16, 2010 Post Count: 799 Status: Offline Project Badges: 
|
Both my Linux PCs run mostly unattended. I dis see this morning one task 'aborted by project' and one which appeared to take some time downloading. Unfortunately, I cannot remember the project name/s - too sleepy to record.
---------------------------------------- crunching, crunching, crunching. AMD Ryzen 5 2600 6-core Processor with Windows 11 64 Pro. AMD Ryzen 7 3700X 8-Core Processor with Windows 11 64 Pro (part time)  |
||
|
|
anhhai
Veteran Cruncher Joined: Mar 22, 2005 Post Count: 839 Status: Offline Project Badges: 
|
Seen only 4 of 6 active sciences. That said, Clean Water is not really of ''loading'' interest because the only thing that's coming down is an instruction to compute x/y/z seed from the table already on the hosts. HCMD2 is also not that hefty, but the exercise would have to be repeated in full come Tuesday cause ''the load could not be simulated'' ;P As for suspending, done that this morning to force those CEP2 tasks ahead that had 5-6-7 hours down, to get the quickest stress relieve, but from there I'll just let it... think the machine will survive without too much trouble. Set up uses 2 physical disks to spread load. --//-- restarting this beta from scratch? Sekerob, did you do the rain dance for me? Because I am 99.99999% guaranteed to get sapphire come the Tuesday's batch of betas.  |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges:
|
About 100min ago I had a temporary upload error for task BETA_op957_00027_17_0, but it resolved itself quickly enough.
02/07/2011 15:11:19 | World Community Grid | Computation for task BETA_op957_00027_17 finished 02/07/2011 15:11:22 | World Community Grid | Started upload of BETA_op957_00027_17_0 02/07/2011 15:11:45 | World Community Grid | Temporarily failed upload of BETA_op957_00027_17_0: connect() failed 02/07/2011 15:11:45 | World Community Grid | Backing off 10 min 17 sec on upload of BETA_op957_00027_17_0 02/07/2011 15:12:00 | | Project communication failed: attempting access to reference site 02/07/2011 15:12:02 | | Internet access OK - project servers may be temporarily down. 02/07/2011 15:20:39 | World Community Grid | update requested by user 02/07/2011 15:20:42 | World Community Grid | Sending scheduler request: Requested by user. 02/07/2011 15:20:42 | World Community Grid | Requesting new tasks for CPU and NVIDIA GPU 02/07/2011 15:20:46 | World Community Grid | Scheduler request completed: got 0 new tasks 02/07/2011 15:20:46 | World Community Grid | No work sent 02/07/2011 15:20:46 | World Community Grid | No work is available for Discovering Dengue Drugs - Together - Phase 2 02/07/2011 15:20:46 | World Community Grid | No work available for the applications you have selected. Please check your settings on the website. 02/07/2011 15:22:04 | World Community Grid | Started upload of BETA_op957_00027_17_0 02/07/2011 15:22:10 | World Community Grid | Finished upload of BETA_op957_00027_17_0 02/07/2011 15:22:13 | World Community Grid | Sending scheduler request: To report completed tasks. 02/07/2011 15:22:13 | World Community Grid | Reporting 1 completed tasks, requesting new tasks for CPU and NVIDIA GPU Slightly odd 1min 18sec delay between update request and upload start. This seems to happen more often with recent Boinc versions. I presently have 144 listed as Valid, though I think some may have been removed by this stage. I need another 15days to take me to red. With 77 IP (including a few 12h tasks) and 30 in PV (about 7days) I should get there, but it depends on which systems run which tasks - the HCC tasks don't take long, and there are not too many FAAH in the field. Guess I could soon be looking at a Red Beta badge, and for the next 3 years! Anyway, going to normalize settings and will try to remember to re-adjust accordingly on Tuesday for any would be re-runs, if we get informed in advance that is. PS. It was odd watching the PV page change (<5min between each): Page: < First 1 2 3 Last > Page: < First 1 2 3 2 Last > Page: < First 1 2 1, 2, 3, 2  |
||
|
|
littlepeaks
Veteran Cruncher USA Joined: Apr 28, 2007 Post Count: 748 Status: Offline Project Badges:
|
I have a questions for the WCG gurus -- since these betas seem to be real science WUs, but downloaded from a different server, are the WUs that are successfully completed considered successfully completed WUs for the science projects, or will they have to be crunched again as non-beta WUs?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It would be good if the repair jobs didn't have such short deadlines. (19 hours and 12 mins)
For some sciences it's fine, but for some it's impossibly short. e.g. HPF2. Most systems seem to do them in 10-16 hrs. Even that makes 19 hrs a very tight deadline. My P4/2.6HT has done several of them, all taking slightly over 22 hours. Now it is doing a repair job with the 19:12 deadline. It'll be at least 3 hours late, so it'll just trigger another WU being sent out unnecessarily. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
In production the scheduler would award extra grace time to the slower devices on top of the standard fraction. Don't know though how predictable the HPF2 times were, but the last HPF2 Beta on me quad is after 7.5 hours on 51% and due in 10... tight. Someone else will get a copy if not making it and it will be canceled if not started before mine finishes and reports. Tasks in deadline exceed report immediately AFAICS.
----------------------------------------Though seeing your concern, certainly I'd not be too worried over the last efficiency squeeze and consider it part of the full robustness test of the added download servers and their integration with the overall system. After the initial hick-ups all started BETA tasks computed here properly to the end [and had about half a dozen that were server aborted for reasons I've not delved into]. Total worth about 12 CPU days. --//-- Afterthought: Maybe in these types of Betas the tasks could be controlled in distribution to only go to those that have opted in for a particular science i.e. if your P4-HT were not opted to receive HPF2 in production, it best not send a beta of HPF2. Would that make sense? [Can see some cons to that suggestion too ;>] edit: bolded for emphasis, general, all science beta's :D [Edit 1 times, last edit by Former Member at Jul 3, 2011 7:22:45 PM] |
||
|
|
|