Message boards :
Number crunching :
Redundant result cancelled by server
Message board moderation
Author | Message |
---|---|
Browny Send message Joined: 22 Sep 99 Posts: 33 Credit: 814,772 RAC: 0 |
568962659 140160055 10 Jul 2007 20:09:37 UTC 11 Jul 2007 5:10:23 UTC Over Redundant result Cancelled by server 0.00 --- --- 568962463 140159975 10 Jul 2007 20:09:17 UTC 11 Jul 2007 5:10:23 UTC Over Redundant result Cancelled by server 0.00 --- --- 568962187 140159898 10 Jul 2007 20:09:17 UTC 11 Jul 2007 7:24:40 UTC Over Redundant result Cancelled by server 0.00 --- --- 568961217 140159649 10 Jul 2007 20:09:17 UTC 11 Jul 2007 5:10:23 UTC Over Redundant result Cancelled by server 0.00 --- --- The files mentioned above were called this morning as "aborted by project"and a few minutes ago they were renamed in "Over Redundant result Cancelled by server. Can somebody tell me what is the meaning of this. With Regards, Browny |
bounty.hunter Send message Joined: 22 Mar 04 Posts: 442 Credit: 459,063 RAC: 0 |
568961217 140159649 10 Jul 2007 20:09:17 UTC 11 Jul 2007 5:10:23 UTC Over Redundant result Cancelled by server 0.00 --- --- Eric has changed the wording to make it clear that it shouldn't be considered an error. I've changed the text displayed for these results on the web site to make it clear that it shouldn't be considered an error. |
ohiomike Send message Joined: 14 Mar 04 Posts: 357 Credit: 650,069 RAC: 0 |
|
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Thank you Eric for the change! Should help clear up some past misunderstandings and avoid some future ones. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Indy Send message Joined: 30 Mar 00 Posts: 9 Credit: 12,750,941 RAC: 0 |
I have 20 WU with "Redundant result" and "Cancelled by server". For example: One of my 20 WU. I received this WU - 6.7.2007, I crunched it in other PC, I copied to my PC with internet and posted it 10.7.2007, BUT I HAVE NO CREDIT ONLY REDUNDANT RESULT !!! It was 34 WU and maybe 2000 credits. :-( :-( :-( |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
I have 20 WU with "Redundant result" and "Cancelled by server". Well there's your problem. You can't use 5.10 series CC's if you're going to 'transplant' the results from the connected host to other machines, or at the very least not let it talk to the project once it DL's work until all of the results are ready to be returned. OOP's I missed that, the stderr say's you're using 5.8.16! Alinator |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
Pilot error! |
Jim-R. Send message Joined: 7 Feb 06 Posts: 1494 Credit: 194,148 RAC: 0 |
I have 20 WU with "Redundant result" and "Cancelled by server". It is not recommended to crunch work on another computer from the one they were issued to. This is what can happen. You'll just have to eat those. You might consider setting up a lan at home to allow the other computer access to the Internet. This way your efforts won't be wasted, and it will be a lot easier than trying to copy files back and forth. Jim Some people plan their life out and look back at the wealth they've had. Others live life day by day and look back at the wealth of experiences and enjoyment they've had. |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
now hangon a second, since when does 5.8.16 respond to a server side abort? "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
now hangon a second, since when does 5.8.16 respond to a server side abort? I didn't think so either, but there it was in black and white!!?? ;-) Alinator |
Jim-R. Send message Joined: 7 Feb 06 Posts: 1494 Credit: 194,148 RAC: 0 |
now hangon a second, since when does 5.8.16 respond to a server side abort? Hmm something is fishy about that one. There's only one other result in, and another one was sent out after this one was canceled. There's been some mention about this happening. Don't know what it is yet. BTW, What version of BOINC are you running on both computers? Same version? It has been reported in Beta but there it seems to have only affected 5.10.x clients Jim Some people plan their life out and look back at the wealth they've had. Others live life day by day and look back at the wealth of experiences and enjoyment they've had. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
Agreed, something's not adding up, but since I never needed to use the transplant method I don't have any first hand experience to go on here. Obviously, theory is breaking down a bit at the moment. ;-) Alinator |
Indy Send message Joined: 30 Mar 00 Posts: 9 Credit: 12,750,941 RAC: 0 |
1. Version 5.8.16 is still recommended 2. I have PC without internet Result: I will crunch other project. Thank you. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
1. Version 5.8.16 is still recommended I think this has less to do with how BOINC behaves, and more to do with how the work was moved between machines, but if you're gone, we'll likely never figure that out. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
1. Version 5.8.16 is still recommended The part which isn't clear is if all the hosts are using 5.8.16, including the one which is used as the 'relay', then how can a 221 happen at all? The only way I can see offhand how it could happen is the relay host would have to be using 5.10 in order to execute the 221. That doesn't explain how the stderr ends up reporting 5.8.16 as the CC version though. However as you said, without any more insight from Indy on the setup and procedures used there's no way to figure just what's happening here. Alinator |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Just A thought, While 5.8.16 doesn't normally honour the server side aborts, will removing (i.e transplanting the incompleted result to another machine) be seen as an acceptance of the server side abort when that machine reports in? This could be verfied by someone running 5.8.16 if they know how to remove a result, and if the result reaches quorum and see if it shows up as a server side abort. If so, this might indicate, as suggested, that if transplanting is needed then the relay host must not be allowed to report in until it has the results back. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Jim-R. Send message Joined: 7 Feb 06 Posts: 1494 Credit: 194,148 RAC: 0 |
Just A thought, I don't believe that this WAS a server abort. At the time the gentleman reported it, and I looked at his results, there was only one result that had reported back outside of his "221" server abort. There was another wu sent out to replace the one of his that aborted! If this had been a real server side abort, why was there only one other result in, and why was another one sent out to replace it? I believe this is more of a validation error than a server abort. Possibly his wu failed validation and the server assigned the wrong code. (This was very near the time when Eric posted that he had just changed the wording of the 221 abort error message.) Jim Some people plan their life out and look back at the wealth they've had. Others live life day by day and look back at the wealth of experiences and enjoyment they've had. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
Just A thought, Thinking about that, 5.8 series CC do support auto resends. So it seems possible that a ghost could get marked as a 221. IOW's when the host sends the work onboard list back to the server, if the missing result was redundant at that point the project wouldn't resend it but 221 it instead. Here's how I envision the relay process working: 1.) First install BOINC on the target unconnected host. 2.) Shutdown BOINC on the slave, and copy the entire BOINC folder to a suitable removable storage device/medium. 3.) Hookup the storage device to the relay host, run BOINC directly from that folder, attach to the desired project, and then DL a suitable amount of work. 4.) Shutdown BOINC on the relay, and then take the storage device/medium back, copy it back onto the slave, and then process it as usual. 5.) When you are getting low on work, or the opportunity arises to get back to the relay, goto item #2 and repeat. The difference here from the original relay session is the relay host should initate a contact session with the project once BOINC is started up on it's own, since as far as it's concerned it now needs to upload, report, and DL new work to refill the cache. I dual boot one of my hosts (the one I'm writing this on as a matter of fact), and run BOINC from a common directory on either 98SE or 2K, depending on which OS I need to be using at the time. This is in essence the same thing as transferring it from one host to another, and I don't have any issues even though it's running 5.10. The only way I can see how we're getting what we see here is if, as you alluded to, Indy is 'trimming' work out of an active crunching host which is acting as the 'agent' for a number of other hosts and then trying to 'graft' that work back into the agent later to report and get new work. I can definitely see how auto resends (and aborts in the case of 5.10) now being enabled would mess that up. So as Ned said, this is most likely a procedural issue in doing the tranplant more than something being 'wrong' with BOINC on either end. The curious thing is Indy's account is showing a run of new host ID's being spawned lately. This is also evidence that a procedural problem exists, since the most common reason for that is for a host to report an lower RPC sequence number than the one the project has on record for the last contact session from the host. This tends to make me think that the agenting scenario is what's going on here, since in the straight forward method I outlined the sequence numbers wouldn't get out of sync. Alinator |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
I have 20 WU with "Redundant result" and "Cancelled by server". Hmm, only 3 minutes from assigned to you till they was reported as aborted... My guess is the web-page is showing the wrong code, and it should really have shown "client detached"... "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
Jim-R. Send message Joined: 7 Feb 06 Posts: 1494 Credit: 194,148 RAC: 0 |
Concerning Indie's situation, this is really weird. This wu was sent out along with one other. WuID Computer Sent 565047427 2581842 4 Jul 2007 20:06:40 UTC Deadline/Received Status 28 Jul 2007 22:16:33 UTC In Progress Unknown New --- This wu is Indies: Wuid Computer Sent 565047428 3553464 6 Jul 2007 13:51:04 UTC Deadline/Received Status 6 Jul 2007 13:53:29 UTC Over Redundant result Cancelled by server 0.00 --- --- Notice the date and time it was sent then "received/cancelled". Approximately two and a half minutes! This next wu was the ONLY one that had been returned at the time of his posting: Notice that it wasn't received until After Indies was canceled. Wuid Computer Sent 565047429 2824910 4 Jul 2007 20:05:52 UTC Deadline/Received Status 7 Jul 2007 2:12:26 UTC Over Success Done 14,106.98 61.71 61.71 And this one had been sent but not received at the time he posted. Notice that it was sent after Indie's was cancelled. Wuid Computer Sent 566463108 1844484 6 Jul 2007 17:03:42 UTC Received Status 11 Jul 2007 23:56:07 UTC Over Success Done 12,276.05 62.06 61.71 Now considering all this, why did the system cancel Indie's wu only two and a half minutes after it was sent? Especially since at the time his was sent, NONE of the other results had been returned? Very weird! The only thing I can think of is some sort of server error. Jim Some people plan their life out and look back at the wealth they've had. Others live life day by day and look back at the wealth of experiences and enjoyment they've had. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.