Message boards :
Number crunching :
How do I Detach without deleting WUs
Message board moderation
Author | Message |
---|---|
MikeW Send message Joined: 7 Apr 04 Posts: 71 Credit: 10,406 RAC: 0 |
I have a PC I want to rebuild from scratch. Right now it has a cache of about 12 SETI and Predictor WUs to clear. I want to detach from the projects when the WUs have been returned, so that the work isn't lost. However, if I disable the network access until the WUs are all crunched, then re-enable it, upload files and report, the client immediately realises that the low water mark has been hit, and downloads a new set of WUs. Can I stop it downloading new work so I can detach cleanly? |
slavko.sk Send message Joined: 27 Jun 00 Posts: 346 Credit: 417,028 RAC: 0 |
Try to set in your preferences chache size to 0 to 0 days. Click here and change Keep enough work on disk to last between field. If you have more preferences, change all. After rebuild of PC set it back to your value. > Can I stop it downloading new work so I can detach cleanly? Watch your credit on WAP Handy/PDA http://slavko.sk/wap/seti |
MikeW Send message Joined: 3 Apr 99 Posts: 16 Credit: 233 RAC: 0 |
Thanks - I'll give it a try. It may take a day or two before I get a result! Giskard - the first telepathic robot. |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
Currently the only way to not get more work, is to either fill-up your hd or increase the disk-preferences so they'll fail with a not enough disk-space for BOINC-message. Setting the cache to 0 will not work, since in this case some default-values will be used instead. |
Arm Send message Joined: 12 Sep 03 Posts: 308 Credit: 15,584,777 RAC: 0 |
|
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
> Another way is to set "Keep enough work on disk to last between" to "0.1" and > "0.1" days. For most of the computers it will lead to the message "No work > from project" I believe. The "no work from project" is just what most see under the best of times! You really want to see "no work (there was work but "something specific to your machine or prefs" )" apprear. Actually, ingside is right.. You need to set the "leave at least" xx GB to be greater than what you have free. In the general prefs.. the home/school/work too wouldn't hurt, but seems they are often ignored. "work for 0 to 0 days" always seems to download atleast 1 WU (1 sec of work requested, for some reason) It would be nice to be able to disable new work on a client basis (by project)... Lets say that you know you will be taking a system down for good, but you have multiple systems. Changes preferences can affect more than one (and probably will). The "by project" option would be nice if you wanted to finish the work you had for a single project (on a single computer), but request no more. For example, it turns out that the mfold app from predictor has a hard time on one of my machines (way too much swapping). Other systems were running fine with mfold. I didn't mind finshing what mfold WU's I had on the "VM chalenged" machine, but wanted no more. Detatch would have dumped what I had in the queue, but was looking for something like "stop_after_send.txt" on a project basis! (a clean way to not leave WU's dangling) |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
> > Another way is to set "Keep enough work on disk to last between" to "0.1" > and > > "0.1" days. For most of the computers it will lead to the message "No > work > > from project" I believe. > > The "no work from project" is just what most see under the best of times! You > really want to see "no work (there was work but "something specific to your > machine or prefs" )" apprear. > > > Actually, ingside is right.. You need to set the "leave at least" xx GB to be > greater than what you have free. In the general prefs.. the home/school/work > too wouldn't hurt, but seems they are often ignored. > > "work for 0 to 0 days" always seems to download atleast 1 WU (1 sec of work > requested, for some reason) > > It would be nice to be able to disable new work on a client basis (by > project)... Lets say that you know you will be taking a system down for good, > but you have multiple systems. Changes preferences can affect more than one > (and probably will). > > The "by project" option would be nice if you wanted to finish the work you had > for a single project (on a single computer), but request no more. For > example, it turns out that the mfold app from predictor has a hard time on one > of my machines (way too much swapping). Other systems were running fine with > mfold. I didn't mind finshing what mfold WU's I had on the "VM chalenged" > machine, but wanted no more. Detatch would have dumped what I had in the > queue, but was looking for something like "stop_after_send.txt" on a project > basis! (a clean way to not leave WU's dangling) > The request has been made to the development team for a detach after all work comploeted. |
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
> The request has been made to the development team for a detach after all work > comploeted. Cool, as long as this means that it's "in limbo" for new work... (no new work when this action has been taken). In other words, not like the -exit_when_idle cmd line option! Also, needed in the cmd line interfaces, and not just the GUI! |
©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.