Thanks for looking into the problem for me, I'm using version 6.10.18
if theres any other specs you need let me know.
thanks again
It looks like that you've simply been unlucky with our recent server woes ( over pressured ) and that your downloads, and the respective wingmen also, weren't satisfactory for a period a few days ago ( HTTP/FTP time out probably ). Not an error from your side that is. Bernd tells me there is a relatively small fraction of WU's put out then ( but still over one hundred of them! ) with 10 or more client errors of this nature!
Sorry, but you've won a lightning strike, or three! With the number of contributors we have someone has to be on the high end of the ( Poisson ) curve when our servers get the wobbles. You're it! :-) :-)
@Gary - since his tasks are still falling over as we speak for those two hosts, and that bad downloads are the likely cause, how should we advise him? What's the best method to recover good input files, with the servers now behaving better? We really don't know what he might/not have that's corrupt. Detach/re-attach ?
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
What's the best method to recover good input files, with the servers now behaving better? We really don't know what he might/not have that's corrupt. Detach/re-attach ?
Error -119 means that it could be his public key his corrupt, which should be solved with a project reset. If that fails, there's always the detach/reattach.
@Gary - since his tasks are still falling over as we speak for those two hosts, and that bad downloads are the likely cause, how should we advise him? What's the best method to recover good input files, with the servers now behaving better? We really don't know what he might/not have that's corrupt. Detach/re-attach ?
I'm not so sure that bad downloads are the problem - particularly for the older Athlon where the comp errors have (with only one exception) nothing to do with downloading. The errors are - exit code -1073741819 (0xc0000005), and apply to ABP2 and S5R6 tasks rather than the new S5GCE. In fact he now has 1 successful S5GCE. It very much looks like a hardware related issue perhaps due to overclocking or inadequate cooling and I'd like to see his answer on that.
For the Phenom, there are more successful results and most show quite a bit of CPU time before the error is flagged. If they were genuine download problems, how could there be part of the computation completed before discovering the missing or corrupt file? I've seen this behaviour before where part way through the computation, one of the input files is suddenly deemed to be corrupt or missing. In my experience, the file being complained about can be quite OK and the real problem is actually heat or overclocking induced.
Also, quite a few of the errors relate to ABP2 and S5R6 tasks which were downloaded well before the onset of S5GCE. It's a bit hard to see why these would be related to overloaded servers or mirrors. Also there are examples where the error is not enough disk space. Could BOINC be deleting or preventing the download of files due to badly set disk preferences? Finally, there is also an example of a task that completed but failed validation. An invalid result could point to a hardware issue.
To help eliminate various possibilities, I'd like to wait until the OP tells us if he is overclocking or not and if he has checked his hardware - CPU cooling, memory testing, etc. I'd also like to know what disk use preferences he has set.
Thanks for looking into the problem for me, I'm using version 6.10.18
if theres any other specs you need let me know.
thanks again
It looks like that you've simply been unlucky with our recent server woes ( over pressured ) and that your downloads, and the respective wingmen also, weren't satisfactory for a period a few days ago ( HTTP/FTP time out probably ). Not an error from your side that is. Bernd tells me there is a relatively small fraction of WU's put out then ( but still over one hundred of them! ) with 10 or more client errors of this nature!
Sorry, but you've won a lightning strike, or three! With the number of contributors we have someone has to be on the high end of the ( Poisson ) curve when our servers get the wobbles. You're it! :-) :-)
Cheers, Mike.
Only 1 machine is overclocked, the AMD Athlon 1ghz is realy 1050 now (I know huge huh :P FSB OC) the AMD 965 and Intel are both stock.
I seen the AMD 1ghz just finished its first WU, either bad luck, or the fact I put the settings from 80% processor usage 80% of the time, to 100% on both.
As for the AMD 965, I'm not sure, seems more hit and miss with WUs, and due to heat I cant run 100%.
I'll let them run see if things clear up, as long as this won't hurt project....will it?
EDIT: Also its not like the 965 overheats, just gets a little warm for my liking at 100%, right now 80% 80% settings I'm running at 52c
This is my gaming PC so I know its running smooth (no I don't game and run BIONIC)
... I put the settings from 80% processor usage 80% of the time, to 100% on both.
This is most likely the reason you now have a successful result. Leave your Athlon on 1GHz and leave the settings (particularly processor usage) on 100%. Hopefully you will see no further failures. For some as yet not properly understood reason, there seems to be a problem with processor usage figures less than 100% on many machines.
Quote:
As for the AMD 965, I'm not sure, seems more hit and miss with WUs, and due to heat I cant run 100%.
To get rid of the comp errors, you should again use 100% processor usage. Heat shouldn't be a problem at stock settings if you are sure your heat sink fins are clean. Remove the fan and take a good look.
Quote:
I'll let them run see if things clear up, as long as this won't hurt project....will it?
No.
Quote:
... right now 80% 80% settings I'm running at 52c
That's a bit frosty these days :-). Maybe if it gets over 65C you could start to worry a bit. If you leave it at 80% processor usage you will probably continue to see comp errors from time to time.
Sorry If I hijacked his thread.
Thanks for trying guys/gals I'll mess with the settings see what happens.
I just now see that the AMD 1ghz ( the PC with the most problems) is running version 6.10.36 the rest are version 6.10.18.
I'll downgrade that later see if it helps.
Thanks again to all who replied, I'll keep an eye on this thread incase someone has some ideas.
RE: Thanks for looking into
)
It looks like that you've simply been unlucky with our recent server woes ( over pressured ) and that your downloads, and the respective wingmen also, weren't satisfactory for a period a few days ago ( HTTP/FTP time out probably ). Not an error from your side that is. Bernd tells me there is a relatively small fraction of WU's put out then ( but still over one hundred of them! ) with 10 or more client errors of this nature!
Sorry, but you've won a lightning strike, or three! With the number of contributors we have someone has to be on the high end of the ( Poisson ) curve when our servers get the wobbles. You're it! :-) :-)
@Gary - since his tasks are still falling over as we speak for those two hosts, and that bad downloads are the likely cause, how should we advise him? What's the best method to recover good input files, with the servers now behaving better? We really don't know what he might/not have that's corrupt. Detach/re-attach ?
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
RE: What's the best method
)
Error -119 means that it could be his public key his corrupt, which should be solved with a project reset. If that fails, there's always the detach/reattach.
RE: @Gary - since his tasks
)
I'm not so sure that bad downloads are the problem - particularly for the older Athlon where the comp errors have (with only one exception) nothing to do with downloading. The errors are - exit code -1073741819 (0xc0000005), and apply to ABP2 and S5R6 tasks rather than the new S5GCE. In fact he now has 1 successful S5GCE. It very much looks like a hardware related issue perhaps due to overclocking or inadequate cooling and I'd like to see his answer on that.
For the Phenom, there are more successful results and most show quite a bit of CPU time before the error is flagged. If they were genuine download problems, how could there be part of the computation completed before discovering the missing or corrupt file? I've seen this behaviour before where part way through the computation, one of the input files is suddenly deemed to be corrupt or missing. In my experience, the file being complained about can be quite OK and the real problem is actually heat or overclocking induced.
Also, quite a few of the errors relate to ABP2 and S5R6 tasks which were downloaded well before the onset of S5GCE. It's a bit hard to see why these would be related to overloaded servers or mirrors. Also there are examples where the error is not enough disk space. Could BOINC be deleting or preventing the download of files due to badly set disk preferences? Finally, there is also an example of a task that completed but failed validation. An invalid result could point to a hardware issue.
To help eliminate various possibilities, I'd like to wait until the OP tells us if he is overclocking or not and if he has checked his hardware - CPU cooling, memory testing, etc. I'd also like to know what disk use preferences he has set.
Cheers,
Gary.
RE: RE: Thanks for
)
Only 1 machine is overclocked, the AMD Athlon 1ghz is realy 1050 now (I know huge huh :P FSB OC) the AMD 965 and Intel are both stock.
I seen the AMD 1ghz just finished its first WU, either bad luck, or the fact I put the settings from 80% processor usage 80% of the time, to 100% on both.
As for the AMD 965, I'm not sure, seems more hit and miss with WUs, and due to heat I cant run 100%.
I'll let them run see if things clear up, as long as this won't hurt project....will it?
EDIT: Also its not like the 965 overheats, just gets a little warm for my liking at 100%, right now 80% 80% settings I'm running at 52c
This is my gaming PC so I know its running smooth (no I don't game and run BIONIC)
RE: ... I put the settings
)
This is most likely the reason you now have a successful result. Leave your Athlon on 1GHz and leave the settings (particularly processor usage) on 100%. Hopefully you will see no further failures. For some as yet not properly understood reason, there seems to be a problem with processor usage figures less than 100% on many machines.
To get rid of the comp errors, you should again use 100% processor usage. Heat shouldn't be a problem at stock settings if you are sure your heat sink fins are clean. Remove the fan and take a good look.
No.
That's a bit frosty these days :-). Maybe if it gets over 65C you could start to worry a bit. If you leave it at 80% processor usage you will probably continue to see comp errors from time to time.
Cheers,
Gary.
Sorry If I hijacked his
)
Sorry If I hijacked his thread.
Thanks for trying guys/gals I'll mess with the settings see what happens.
I just now see that the AMD 1ghz ( the PC with the most problems) is running version 6.10.36 the rest are version 6.10.18.
I'll downgrade that later see if it helps.
Thanks again to all who replied, I'll keep an eye on this thread incase someone has some ideas.
ZZUBY