NBP 6.30B3 Build 1718 Program "thrashing"

[bertm]

First, I have tried 6.20 and experience the same results.

I select "Update All headers", when NBP gets to a few major groups (a.b.teevee, a.b.multimedia for example) it lists headers to DL, but the download count never increases from zero. I let it run all night and it just "churned." Every time I close the program I need to wait about five minutes before NBP exits from task manager. Each time I open to "Update All Headers" these same groups get hung up, but the headers to download grows- telling me no updating of headers has occurred.

Is there any solution beyond re-DLing all headers from these troubled groups?

PS I ran a chkdsk on the data drive. There were no problems.

[Quade]

Look down at the "Cache" line. Do you see any numbers in parens "(NN/MM)"? Is so, these are records waiting to be imported.

In the options, I'd suggest checking "Background Update" if it's not currently checked.

[bertm]

Still "Thrashing" - Cache stuck at 75/100. I set to background update as well.

[Quade]

The 75/100 thing isn't what I mean. If you don't have numbers in parens, you don't have any waiting to import.

I might exit, then delete "Downloads.db3" to wipe out the download list. It's not really clear what's going on. Might be good to do a clean sweep. Might kill "Autopar2.db3" while you're at it.

[bertm]

I watched the Cache number (xx) go from 48 to 0 as I updated all headers.

I deleted autopar.db3 and downloads.db3 but nothing changed.

I deleted the header DL for a couple of groups and selected DL all headers for each but nothing changed (except the number of headers to DL, obviously).

I plan to purge to display age and compact DB next.

[Quade]

I watched the Cache number (xx) go from 48 to 0 as I updated all headers.

Once this is done, there should be no more disk IO (well, unless you're downloading). I'm not really getting a feeling for what's happening. If it's downloading headers, particularly at high speeds, the disk drive is going to get worked pretty hard. Not sure I'd call that thrashing though. The data has to go somewhere. When I'm downloading headers at 400 Mbps that means I have to write to the disk at some percentage of 400 Mbps. In something like picture groups, it's going to be pretty heavy writing. In groups with larger files, there's actually less to write to disk.

At this point I'm not really hearing a problem. How fast are the headers downloading?

[bertm]

Right now I have been purging my multimedia folder, Which includes a.b.mm, a.b.teevee, etc.

Previously, I updated all headers, the parentheses went away but about 6 or 7 groups with 0/xxxxxx headers to DL never updated, nor went away.

Even when I removed these inactive groups from the DL list nothing changed - the disk activity remained.

For now I will let the purge run over night, if necessary, and see what happens.

BTW I use 32 bit XP - are there any capacity issues I should be concerned with - days storage, etc.

[Quade]

Sounds like you have 4-5 different things going on at once. You need to simplify this down before I can understand it. If you ask Newsbin to beat the crap out of your disk drive's, it'll do it. It's sort of up to you to see how much your machine can take and not work it too hard.

Perfect example, I probably wouldn't download headers and try to purge groups at the same time. You're going to run out of disk bandwidth in short order.

[bertm]

Purging to display age seems to have done the trick. Maybe it re-indexed the files?

Thanks

Just a follow-up:

I had my wishlist loaded - but now they are gone. No biggie - just mentioned in case that may have been a contributing factor.

What I think may have contributed to my problems are:

I had the drive set for intelligent defrag - it's now off.

I had to re-DL headers after a unsuccessful change of newsgroup providers. I think I may have overdone the DL 10,000,000 older headers to backfill what was missing.

Let me know if these may have contributed to the "churning."

[DThor]

Wow, 'intelligent defrag '. That sure sounds like a huge part of the problem. I have a policy to never trust software with the words 'intelligent ' or 'smart' in the title. Seriously.

There's lots of variants on this name out there, so it's hard to tell, but if something starts up a defrag process while an application is in the middle of doing mass inserts into a database, this definitely qualifies as a Bad Thing.

DT

[bertm]

In retrospect I agree!