I have a machine running at a location I only get to every 2-3 months or longer. I figured CP would be ideal for a situation such as this. But now that the wu may finish in the next month or so I wont be able to download a replacement on my next trip cause it will be more than 10 days. Which of course seems dumb with wu\'s woth 6 month or 1 year deadline.

I think I know the answer to this e.g. the BONIC server just wont allow the person owning the host to make such a high level decesion.

Can I still get the non Bonic version of the program so I cam keep all my machine goings? As several times a year I may be unable to get to all of my hosts for 30 days or so.

---

The probable answer to your question is that it was set at a low level to prevent people from caching vast number of WUs in SETI, which also has comparatively short deadlines so that there wouldn\'t be much point in getting more for the machine anyway. I agree that this isn\'t really relevant to CPDN, which is probably the only suitable project in your situation.

I can\'t think of an easy solution that would met the needs of all projects, and of course the setting is in general preferences and so affects every project. Perhaps someone can suggest something?

EDIT OK, you have now answered your original question. Yes you can use classic, but I don\'t think it would help, as it was not designed to allow any caching of units. It calls for a new one only when it finishes the first.

---

You can manually edit your global_perfs.xml file and insert a larger value.

I personally support a more aggressive limit to the queue size, however it should be a project set limit rather than hard coded into the system. My idea is that the project would automatically reduce the queue to the project\'s max even if it was set larger at another project. This would have almost no impact here at CPDN since the max here would probably be 6 months. However at a project like PP@H the max would likely be 3 or 4 days. Basically the limit should be 50% to 75% of a normal deadline.

---

BOINC WIKI

BOINCing since 2002/12/8

Can\'t you do that by remote hosting?

---

BOINC Wiki

if you just want to force it do download a 2nd model you can suspend the 1st model, it should download and start a new model then, unsuspend the 1st (exit and restart boinc to get the 2nd model out of memory if thats a problem) and be happy :)
no idea though at what time a model is considered \"lost\" by the server if it doesnt hear from it for a such long time.

It\'s \"lost\" after about 6 weeks.