newsbin.com

Yesterday I downloaded headers of a couple of groups, such as alt.binaries.cores (about 2.75 billion posts) from two servers at once (giganews, easynews) and over 40 of such files as:

Storage.db3-mj0E27E49D7
Storage.db3-mjC94B8394B
....
still remain even when the background processing ended and the cache reached 0.
Even after I restarted NewsBin.
Even after I ran additionally "Download latest"
Do they indicate some corruption or I can just delete them?

Until now, I always downloaded headers from giganews only (at least V6 headers) and I see only one such file in all the headers of various groups I have.
In my experience those files disappear when the cache reaches 0.
But not in this case.

They're temporary files generated by Sqlite. I'd say delete and keep an eye on things. Suggests maybe Newsbin crashed or exited while Sqlite was in the middle of things. The latest beta's should never generate these files at all. Are the date's of the files current?

For example, a.b.cores - I never had the headers of this group, downloaded it for the first time a couple of days ago with NewsBin 6.50 beta 16 and those Storage.db3-mj... files were created, they are new.
Also, newsbin did not crash.
I remember that usually those files disappear when cache goes to 0.
Oh, there probably was a pause when the drive with headers became full. Maybe that caused it?
But are those fresh headers now corrupt?

I was downloading at the same time a.b.town, so I noticed that a.b.town was downloaded in a single pass, but for a.b.cores at some point the count changed and the progress bar began from zero again.
However, those Storage.db3-mj files are left for a.b.town too.

a.b.cores - almost 42 GB of headers, 74 Storage.db3-mj files remaining, a.b.town - 25 GB of headers, 42 files remaining

If you run out of disk space, bad things happen. I recommend not letting it happen.

I doubt the database is corrupt though, you might be missing some posts.

Should I set headeroverlap to something?
It seems like NewsBin should know better what headers to download not to leave any gaps, but overlap probably does not create any duplications, just downloads some headers perhaps unnecessarily again?

Your guess is as good as mine. I don't test running out of disk space so, I have no idea what happens when it happens. Everything might be fine, then again it might not.

That's why I recommend not running out of space.