I'm just catching up with progress whilst I've been away and it's good to see everything looking fairly normal. So, two completed and validated EAH results now. Your total credit is now 140 which looks very good.
I'm assuming that your tone of disappointment was due to the fact that you have a couple of client errors. Michael has suggested there is no problem with those and he is correct. Here is what I think is happening which is causing these client errors. Unfortunately there may not be much you can do about it.
Your second successful result finished at 19 Oct 2005 9:07:02 UTC. The name of the result file was l1_1097.5__1097.7_0.1_T07_S4lC_0 and you would have been working on a large data file called l1_1097.5 which had been downloaded 10 days earlier. Normally, when one result finishes, another is "sliced" from the same large data file and so there is no fresh large data file to be downloaded. In your case however, things are complicated by two particular factors.
Firstly, because your computer is very slow and there are many other people crunching on the same data, by the time a new set of work is needed for your computer, others have already "used up" all the remaining data. So the server tells your BOINC client it must discard the now "used up" large data file and get a new one. The "client error" result tells us the name of the new data file as being w1_1102.5. So just before your second result was completed on 19 Oct, BOINC was instructed to get the new data file, w1_1102.5.
This leads to the second complication. Your internet connection is very unstable and suffers frequent disconnects. Such a disconnect could cause the transfer of the large data file to fail part way through. I'm guessing that your local BOINC "errors out" the next listed set of work because of the download failure. When your connection is re-established, BOINC would continue the data download from where it left off previously and eventually the whole file would be on your computer and the next result could crunch normally.
I'm guessing the above scenario and the only concern is that I would have thought that since BOINC is designed to be fault tolerant, it shouldn't have immediately "errored out" the result just because of the connection loss. BOINC would expect to be able to continue with the download once the link was reconnected. So I'm wondering if you might have taken some action actively to abort the transfer rather than the problem just being caused by a flakey connection. In that case I could more easily understand the client error.
Anyways, it doesn't really matter as the new file w1_1102.5 was eventually downloaded and you are now crunching on result w1_1102.5__1102.5_0.1_T03_S4hC_2. In about 7 days time, the whole process is likely to repeat as the current data file may well be used up by others by then. However this really shouldn't concern you too much.
Whew, am I glad you're back! Your short departure created a partial vacuum on the help reesponse. Stick and I have been trying to cover the gap. I don't think we've mucked anything up too bad and we've probably done a passable job quantity-wise, passing out a lot of bandaids, and the "chief surgeons" have done admirably in picking up where our knowledge left off, but he and I lack your in-depth knowledge, that much has been evident.
I was distressed by the tone of dismay in Ariane's most recent post, and checked the Seti boards yesterday evening, to find that she hadn't been on there since about the same time as she was here. However, I checked over there just now and was happy to discover that she's had some activity about 8 hours ago, so she hasn't dropped off the end of the world. Still, there's that "The miracle has ended".
Your friend,
Michael
microcraft
"The arc of history is long, but it bends toward justice" - MLK
It's true, these few days , my internet connection was horrible, given i use internet cards , ( i work by dial up 0, normally the maximum of speed i had always was between 48 to 50 Kbps , but mostly through the past week this speed felt to 14 and 10 kbps and just now i'm connected with 14 kbps and each second i'm disconnected ,it's incredible.Today i called my card server and asked about this speed trouble, he told me that the problem has been solved, but now i see that it is not solved yet. Anyway i hope everything would become normal very soon and i would be able to crunch like the past and The Miracle would retake its ordinary train .
Whew, am I glad you're back! Your short departure created a partial vacuum on the help reesponse. Stick and I have been trying to cover the gap .....
And you've both been doing a mighty fine job. I've just done a bit of a catch-up on what's been happening while I've been away and I must commend you both for your dedication to helping others. I'll try and do a bit each day to help as well from now on. However both of you have done extremely well and the project is fortunate to have you and many other fine helpers.
.... Today i called my card server and asked about this speed trouble, he told me that the problem has been solved, but now i see that it is not solved yet. Anyway i hope everything would become normal very soon and i would be able to crunch like the past and The Miracle would retake its ordinary train .
Cheers
Hi Ariane,
You might have to keep nagging your ISP to get better service and a more satisfactory browsing experience. I imagine it must be very frustrating if you are trying to read and respond to messages. Fortunately, it should have little impact on the majority of your crunching. You can be offline for most of the time and just allow BOINC to connect when it needs to upload results or download new work. This is a very predictable time for EAH. I believe we left you with a very small "connect to network" interval of about 0.05 days. I don't remember the precise number. The idea was to delay until the very last moment, the downloading of the next work unit. Now that you have mentioned speeds of around 10-14 kbps, I can see a problem if a large data file is needed. It may take more than 1 hour to download the large data file so perhaps we need to start the download a little earlier so that it will have completed before the currently crunching result actually finishes.
Please check your account on the website and tell us the value set for the "connect to network" interval. I think we should make it at least 0.08 days or perhaps a bit larger so that there is enough time to get new data before it is actually needed.
OK, thanks for that. 0.03 days is 0.72 hours. A large data file is around 6+MB and even if you had a sustained speed of 10-14kbps, 0.72 hours would not be enough time to complete the download before BOINC needed the file for the start of the next result. This would be the reason why you are having a result give a client error when it tries to start.
So, all you need to do is edit your preferences on the website to change the value from 0.03 to 0.08. Then save the change on the website and you should not have this problem next time. Whilst BOINC is able to cope with the problem with minimal interruption, we may as well fix it properly.
Yes everything is OK , I got my new credits from Seti 2 days ago and it started its third project.
Actually, Seti is on the 4th result. You have already finished three!! :).
Quote:
EAH is running well , the Progress just now is 37.59 %
Seti's Progress is 21.18%
You have just over 8 more days for the EAH deadline, so that should be enough. Because Seti is much shorter, you may finish your current Seti result before you finish EAH. They might be quite close together near the end. I'll ask again in a few days to see how things are progressing. It's always a pleasure to talk with you. Take care!!
Hi Ariane and Michael, I'm
)
Hi Ariane and Michael,
I'm just catching up with progress whilst I've been away and it's good to see everything looking fairly normal. So, two completed and validated EAH results now. Your total credit is now 140 which looks very good.
I'm assuming that your tone of disappointment was due to the fact that you have a couple of client errors. Michael has suggested there is no problem with those and he is correct. Here is what I think is happening which is causing these client errors. Unfortunately there may not be much you can do about it.
Your second successful result finished at 19 Oct 2005 9:07:02 UTC. The name of the result file was l1_1097.5__1097.7_0.1_T07_S4lC_0 and you would have been working on a large data file called l1_1097.5 which had been downloaded 10 days earlier. Normally, when one result finishes, another is "sliced" from the same large data file and so there is no fresh large data file to be downloaded. In your case however, things are complicated by two particular factors.
Firstly, because your computer is very slow and there are many other people crunching on the same data, by the time a new set of work is needed for your computer, others have already "used up" all the remaining data. So the server tells your BOINC client it must discard the now "used up" large data file and get a new one. The "client error" result tells us the name of the new data file as being w1_1102.5. So just before your second result was completed on 19 Oct, BOINC was instructed to get the new data file, w1_1102.5.
This leads to the second complication. Your internet connection is very unstable and suffers frequent disconnects. Such a disconnect could cause the transfer of the large data file to fail part way through. I'm guessing that your local BOINC "errors out" the next listed set of work because of the download failure. When your connection is re-established, BOINC would continue the data download from where it left off previously and eventually the whole file would be on your computer and the next result could crunch normally.
I'm guessing the above scenario and the only concern is that I would have thought that since BOINC is designed to be fault tolerant, it shouldn't have immediately "errored out" the result just because of the connection loss. BOINC would expect to be able to continue with the download once the link was reconnected. So I'm wondering if you might have taken some action actively to abort the transfer rather than the problem just being caused by a flakey connection. In that case I could more easily understand the client error.
Anyways, it doesn't really matter as the new file w1_1102.5 was eventually downloaded and you are now crunching on result w1_1102.5__1102.5_0.1_T03_S4hC_2. In about 7 days time, the whole process is likely to repeat as the current data file may well be used up by others by then. However this really shouldn't concern you too much.
Cheers,
Gary.
RE: Hi Ariane and
)
Hi Gary,
Whew, am I glad you're back! Your short departure created a partial vacuum on the help reesponse. Stick and I have been trying to cover the gap. I don't think we've mucked anything up too bad and we've probably done a passable job quantity-wise, passing out a lot of bandaids, and the "chief surgeons" have done admirably in picking up where our knowledge left off, but he and I lack your in-depth knowledge, that much has been evident.
I was distressed by the tone of dismay in Ariane's most recent post, and checked the Seti boards yesterday evening, to find that she hadn't been on there since about the same time as she was here. However, I checked over there just now and was happy to discover that she's had some activity about 8 hours ago, so she hasn't dropped off the end of the world. Still, there's that "The miracle has ended".
Your friend,
Michael
microcraft
"The arc of history is long, but it bends toward justice" - MLK
GARY ! My dear friend :
)
GARY ! My dear friend :
It's true, these few days , my internet connection was horrible, given i use internet cards , ( i work by dial up 0, normally the maximum of speed i had always was between 48 to 50 Kbps , but mostly through the past week this speed felt to 14 and 10 kbps and just now i'm connected with 14 kbps and each second i'm disconnected ,it's incredible.Today i called my card server and asked about this speed trouble, he told me that the problem has been solved, but now i see that it is not solved yet. Anyway i hope everything would become normal very soon and i would be able to crunch like the past and The Miracle would retake its ordinary train .
Cheers
Ariane
RE: Whew, am I glad you're
)
And you've both been doing a mighty fine job. I've just done a bit of a catch-up on what's been happening while I've been away and I must commend you both for your dedication to helping others. I'll try and do a bit each day to help as well from now on. However both of you have done extremely well and the project is fortunate to have you and many other fine helpers.
Cheers,
Gary.
RE: .... Today i called
)
Hi Ariane,
You might have to keep nagging your ISP to get better service and a more satisfactory browsing experience. I imagine it must be very frustrating if you are trying to read and respond to messages. Fortunately, it should have little impact on the majority of your crunching. You can be offline for most of the time and just allow BOINC to connect when it needs to upload results or download new work. This is a very predictable time for EAH. I believe we left you with a very small "connect to network" interval of about 0.05 days. I don't remember the precise number. The idea was to delay until the very last moment, the downloading of the next work unit. Now that you have mentioned speeds of around 10-14 kbps, I can see a problem if a large data file is needed. It may take more than 1 hour to download the large data file so perhaps we need to start the download a little earlier so that it will have completed before the currently crunching result actually finishes.
Please check your account on the website and tell us the value set for the "connect to network" interval. I think we should make it at least 0.08 days or perhaps a bit larger so that there is enough time to get new data before it is actually needed.
Cheers,
Gary.
Dear Gary , The Value
)
Dear Gary ,
The Value set for "connect to network" is 0.03
Ariane
RE: The Value set for
)
OK, thanks for that. 0.03 days is 0.72 hours. A large data file is around 6+MB and even if you had a sustained speed of 10-14kbps, 0.72 hours would not be enough time to complete the download before BOINC needed the file for the start of the next result. This would be the reason why you are having a result give a client error when it tries to start.
So, all you need to do is edit your preferences on the website to change the value from 0.03 to 0.08. Then save the change on the website and you should not have this problem next time. Whilst BOINC is able to cope with the problem with minimal interruption, we may as well fix it properly.
Cheers,
Gary.
Hi Ariane, Did you get the
)
Hi Ariane,
Did you get the value changed OK?
How is your next EAH result going? I imagine it must be around 50% complete by now?
Cheers,
Gary.
RE: Hi Ariane, Did you get
)
Hi GARY, my dear friend (Hugs)
Yes everything is OK , I got my new credits from Seti 2 days ago and it started its third project.
EAH is running well , the Progress just now is 37.59 %
Seti's Progress is 21.18%
Cheers
Ariane
RE: Yes everything is
)
Actually, Seti is on the 4th result. You have already finished three!! :).
You have just over 8 more days for the EAH deadline, so that should be enough. Because Seti is much shorter, you may finish your current Seti result before you finish EAH. They might be quite close together near the end. I'll ask again in a few days to see how things are progressing. It's always a pleasure to talk with you. Take care!!
Cheers,
Gary.