Hi! Joris Trooster wrote: > > > > nntpcach@fozzy<438>: gdb -c core ../sbin/nntpcached > > > [...] > > > #0 make_vm_proc (state=nc_none, client=8, name=0x3cdf0 "update") > > > at nntpcache.c:496 > > > 496 Stats->client_connects++; > > > > > > I get the same result, when I try to open a connection: > > > > > > Feb 4 16:24:07 fozzy nntpcached[24098]: nntpcache.c:399: starting > > > client task > > > Feb 4 16:24:07 fozzy nntpcache-client[24105]: client task awakening > > > Firstly, what OS/architecture are you running on? > > I have the same problems. nntpcached starts up without any problems. > However connecting to the nntp or http port results in a core dump. > Nntpcached keeps running, but every connect to the nntp of http port > results in a new core dump. > > Like the other reports on the list I'm using Solaris 2.5 > I have something similar on my Linux Alpha system. Some debugging revealed that the forked nntpcache process dies due to a SIGSEGV in print_active, processing a "LIST" command. I could track it down to a problem with the mmaped files, as the pointer to the newsgroup_index pointer "Ni" has a completely illegal value. It _could_ be a problem with different sizes of data types on the Alpha architecture (64bit): a short glance at the sources showed a mix-up of function-argument types, but I haven't fully debugged it yet. The SIGSEGV went away after deleting all the cache.* files and restarting nntpcached. I don't know if this is the problem you have on Solaris, but maybe it's worth a try. - andreas -- Andreas Haumer | email: andreas@xss.co.at | PGP key available *x Software + Systeme | phone: +43.1.6001508 | on request. Buchengasse 67/8 | +43.664.3004449 | A-1100 Vienna, Austria | fax: +43.1.6001507 |
