Re-try large set wit incompletes = thrash & crash

[darkfalz]

Astraweb is having terrible completion lately. As a result I am often having large sets with too many incompletes for the PAR2 set to repair.

Sometimes if I retry at a later time, some of those missing articles will have populated. Unfortunately with most large sets, this causes Newsbin to lock up with constant, heavy disk activity and continuously growing memory usage (the constant thrashing may be the result of pagefile swapping). The only solution is to kill it (which sometimes works) or to force shutdown the PC.

[DThor]

Not sure what you're asking, and without information about exactly how hard you're pushing your system and what sort of system it is, there's not a lot to say other than don't push your system so hard. For example, don't mix header download and insertion(which goes on for a while) with par repair. The growing memory usage is odd, sure memory can climb, but of course I don't know how much you've got, or what else you're running at the same time. As far as page swap goes, yeah, my experience with windows is once you get into that world, you get an unstable and mostly useless system.

DT

[Quade]

What version?

[darkfalz]

I haven't noticed this behaviour lately, maybe it was fixed in some version.

Basically what would happen (in a list with large numbers of incompletes) I would select all, and then hit "Re-try" - this would be followed by increasing disk activity and memory usage until Newsbin became completely unresponsive (though in the background, memory usage and disk use would continue).