JOHNSON@NORTHEASTERN.EDU (I am only an egg.) (07/19/89)
Have a problem here.
I recently asked my news feed to open the flood gates and give me a
full feed. He did. I am getting this from two different nodes at MIT.
I found that I had two incoming NNTPs running, one from each node, and
nothing happening to NNTP.BATCH. I looked and both of them have
NEWS.ITEMS and NEWS.GROUPS open. The last time I looked, neither of
these was opened shared. Is it possible I could have a dead lock here
and that's why NNTP.BATCH isn't getting bigger? If so, will putting
shared bits in the RMS open for these files fix the problem?
HELP.
Chris Johnson
ANU-NEWSer
Northeastern U.
gih900@UUNET.UU.NET (Geoff Huston) (07/20/89)
Chris Johnson writes :-
> I recently asked my news feed to open the flood gates and give me a
>full feed. He did. I am getting this from two different nodes at MIT.
>I found that I had two incoming NNTPs running, one from each node, and
>nothing happening to NNTP.BATCH. I looked and both of them have
>NEWS.ITEMS and NEWS.GROUPS open. The last time I looked, neither of
>these was opened shared.
In both NNTP_SERVER.C and NNTP_FEED.C these files are open with full shared
access:
static
open_server_files()
{
...
itmfab.fab$b_shr = FAB$M_SHRDEL | FAB$M_SHRGET | FAB$M_SHRPUT |
FAB$M_SHRUPD;
...
grpfab.fab$b_shr = FAB$M_SHRDEL | FAB$M_SHRGET | FAB$M_SHRPUT |
FAB$M_SHRUPD;
if (!(sys$open(&itmfab) & 1)) return(0);
if (!(sys$connect(&itmrab) & 1)) return(0);
if (!(sys$open(&grpfab) & 1)) return(0);
if (!(sys$connect(&grprab) & 1)) return(0);
return(1);
}
> Is it possible I could have a dead lock here
>and that's why NNTP.BATCH isn't getting bigger? If so, will putting
>shared bits in the RMS open for these files fix the problem?
I'll need more information to provide a more helpful response - I'm a bit
suprised if these files have really been opened non-shared
Geoff Huston
gih900@csc.anu.oz.au