dws@cseg.uucp (David W. Summers) (11/23/88)
I got the NNTP Patch 3 a few weeks (months?) back and applied it and every- thing seems to be working great. I really like the "Out of Space" part of the patch which checks to see whether there is enough space on the disk to enter the article into the system. However, I have one question which I have not seen anybody address: Why stop at just checking the number of disk blocks? Why not also check the number of I-NODES available? That information is included in the structure returned from the call that checks the number of available disk blocks. I couldn't see any reason not to put that check into the same place as the disk block check, and modified 'misc.c' (I think) to do that. It seems to have caused no problems and greatly helped on my machine where the number of I-NODES gets close to 0 quite often. Now whenever the number of I-NODES goes below some minimum (500?) the NNTP daemon refuses to transfer (is that the correct terminology?) any more articles (just like it does when the number of disk blocks gets below some minumum). Like I said, it seems to be working great. If anyone is interested I can send the diffs. Is there some reason that I'm not thinking about where that could get me into trouble? If not, then I would like to see that in the next patch to NNTP (the next time one comes out). Comments? - David Summers (dws@cseg.uucp)