STABLE-20040123 + OVDB: nnrpd dies on signal 11
rra at stanford.edu
Mon May 17 06:39:56 UTC 2004
Heath Kehoe <hakehoe at avalon.net> writes:
> The coredumps are probably caused by a stack overflow, because of a loop
> in the exit (process calls exit(), there's an exit handler that calls
> ovdb_close(), which calls the csend() function which calls exit() when
> the write fails).
> It could be that the nnrpd is just trying to exit normally.
This has now been fixed by preventing csend from ever calling exit, since
that seems wrong anyway (at the least, it should let nnrpd call its own
exit routine so that it can log statistics). The solution isn't ideal
since we may still confuse client errors with missing articles, but it's a
bit better. A full fix will require better error reporting from the
Russ Allbery (rra at stanford.edu) <http://www.eyrie.org/~eagle/>
Please send questions to the list rather than mailing me directly.
<http://www.eyrie.org/~eagle/faqs/questions.html> explains why.
More information about the inn-workers