After I saw the validated tasks I reset the project.......
Wasn't sure if that was clearing what-ever-was-wrong......
So deleting the project.....
Than re-attached the project and, like Bernd said, get jobs re-assignd.
After I saw the validated tasks I reset the project.......
Wasn't sure if that was clearing what-ever-was-wrong......
So deleting the project.....
Than re-attached the project and, like Bernd said, get jobs re-assignd.
I don't have time to follow BOINC development so I'm quite `trapped in the past' and remember some of the 'gotchas' from a long time ago :-). I tend to fix things the hard way (ie manual editing of files) rather than relying on the simple way of clicking a button or two .... mainly because i'm not sure what the current version of the code behind that button is actually going to do :-).
For example, in your case you must have 'reset' the project, then 'removed' the project, then 'added' it back again. Is that correct? Because I don't know anything about the code, I'd be concerned that doing that sequence would remove your current hostID and get you a new one. Doing that could cause a removal of your hosts 'ownership' of the 'lost tasks' sitting on the server. Now I hasten to add that those concerns are obviously unfounded, since you have retained your hostID and you now have all the lost tasks. I'm sure that in my dim dark past somewhere, there was a rather different outcome to doing something similar to what you have done.
I'm very pleased that it all got sorted for you and perhaps I'll profit from your experience and learn to trust the 'one click fixes' a bit more :-).
Quote:
Case closed...(?)
From your perspective, certainly. I'm sure Bernd will be trying to figure out how this happened in the first place.
@Bernd: Thanks for the explanation. I didn't know you could process scheduler requests by email!! That's quite amusing :-). I knew the 'resend lost results' feature would have been responsible for all the new tasks but was puzzled by how that could happen if the 'old' host couldn't actually talk to the scheduler itself. I wonder if I send the lady in my Sat Nav system a grovelling email, she might start talking to me again too :-).
For example, in your case you must have 'reset' the project, then 'removed' the project, then 'added' it back again. Is that correct? Because I don't know anything about the code, I'd be concerned that doing that sequence would remove your current hostID and get you a new one. Doing that could cause a removal of your hosts 'ownership' of the 'lost tasks' sitting on the server. Now I hasten to add that those concerns are obviously unfounded, since you have retained your hostID and you now have all the lost tasks. I'm sure that in my dim dark past somewhere, there was a rather different outcome to doing something similar to what you have done.
I'm very pleased that it all got sorted for you and perhaps I'll profit from your experience and learn to trust the 'one click fixes' a bit more :-).
@Bernd: Thanks for the explanation. I didn't know you could process scheduler requests by email!! That's quite amusing :-). I knew the 'resend lost results' feature would have been responsible for all the new tasks but was puzzled by how that could happen if the 'old' host couldn't actually talk to the scheduler itself. I wonder if I send the lady in my Sat Nav system a grovelling email, she might start talking to me again too :-).
Well,,,,,the workunits sat on that laptop quite a while before they get validated (2 weeks)
So I had to practice my patience while you guys had to do the hard work by sorting out the problems serverside.
But when I read the "all's well" messages, I got worried...!
After I saw the 73 results on the account page and the contactdate of that laptop wasn't jan 7 anymore I got confident and did the 'reset, remove and add project'.
Worst case would be that I didn't get these "send" workunits but someone else....
There's probably a process that re-assign the ID to the identical hardware within a certain timeframe. (?)
I think my re-attach to the project was within a minute.
About Miss Sat Nav; Do not go into a discussion with her about directions!
(you might get punished with a deadly silence)
You might get even to change the voice?
There's probably a process that re-assign the ID to the identical hardware within a certain timeframe. (?)
No, not timeframe. The process is bound to the 'host cross-project id' (CPID), which is derived from the hardware address (MAC) of your main ethernet interface (and some other info related to your E@H account). Without certain measurements the host CPID will even be the same after a complete removal and re-installation of the BOINC Client on a given machine.
Thanks for helping me
)
Thanks for helping me out.
After I saw the validated tasks I reset the project.......
Wasn't sure if that was clearing what-ever-was-wrong......
So deleting the project.....
Than re-attached the project and, like Bernd said, get jobs re-assignd.
Now crunching like nothing happend.......
Case closed...(?)
RE: After I saw the
)
I don't have time to follow BOINC development so I'm quite `trapped in the past' and remember some of the 'gotchas' from a long time ago :-). I tend to fix things the hard way (ie manual editing of files) rather than relying on the simple way of clicking a button or two .... mainly because i'm not sure what the current version of the code behind that button is actually going to do :-).
For example, in your case you must have 'reset' the project, then 'removed' the project, then 'added' it back again. Is that correct? Because I don't know anything about the code, I'd be concerned that doing that sequence would remove your current hostID and get you a new one. Doing that could cause a removal of your hosts 'ownership' of the 'lost tasks' sitting on the server. Now I hasten to add that those concerns are obviously unfounded, since you have retained your hostID and you now have all the lost tasks. I'm sure that in my dim dark past somewhere, there was a rather different outcome to doing something similar to what you have done.
I'm very pleased that it all got sorted for you and perhaps I'll profit from your experience and learn to trust the 'one click fixes' a bit more :-).
From your perspective, certainly. I'm sure Bernd will be trying to figure out how this happened in the first place.
@Bernd: Thanks for the explanation. I didn't know you could process scheduler requests by email!! That's quite amusing :-). I knew the 'resend lost results' feature would have been responsible for all the new tasks but was puzzled by how that could happen if the 'old' host couldn't actually talk to the scheduler itself. I wonder if I send the lady in my Sat Nav system a grovelling email, she might start talking to me again too :-).
Cheers,
Gary.
RE: For example, in your
)
Well,,,,,the workunits sat on that laptop quite a while before they get validated (2 weeks)
So I had to practice my patience while you guys had to do the hard work by sorting out the problems serverside.
But when I read the "all's well" messages, I got worried...!
After I saw the 73 results on the account page and the contactdate of that laptop wasn't jan 7 anymore I got confident and did the 'reset, remove and add project'.
Worst case would be that I didn't get these "send" workunits but someone else....
There's probably a process that re-assign the ID to the identical hardware within a certain timeframe. (?)
I think my re-attach to the project was within a minute.
About Miss Sat Nav; Do not go into a discussion with her about directions!
(you might get punished with a deadly silence)
You might get even to change the voice?
Aad
RE: There's probably a
)
No, not timeframe. The process is bound to the 'host cross-project id' (CPID), which is derived from the hardware address (MAC) of your main ethernet interface (and some other info related to your E@H account). Without certain measurements the host CPID will even be the same after a complete removal and re-installation of the BOINC Client on a given machine.
BM
BM