Red Hat Bugzilla – Bug 6529
if linuxconf can't understand /etc/named.conf it core dumps
Last modified: 2008-05-01 11:37:52 EDT
On our new RH61 server we have a large /etc/named.conf (around 3500 lines). Most of it has been written by hand and the original was converted from 4.9.x format. When we try and run linuxconf, it coredumps. If we remove the /etc/named.conf file and run linuxconf it works. But now that I'm filling out these bugsheet I can't replicate the problem. I still have the core file however (http://users.interact.net.au/~tom/linuxconf-rh61-core).
Can you send me your named.conf so I can look at it. The problem might be related to a zone file though.
I've sent off my named.conf to jack, it's too big (and public) to go here.
Linuxconf 1.16r10 has some fixes for the DNS which are solving this cases. The listen-on directives was improperly parsed when the optional port directive was used. Also, the allow-transfer did not accept a network specification with the /subnet. It was only expecting a host. It should be out today (nov 30). Check it out.
We are currently working on an errata release based on 1.16r10, and we'll make a release after sufficient testing based on 1.16r10 or later.
Linuxconf-1.16r10 will take a total garbage /etc/named.conf file and try it's darndest to parse it, but no core dumps :-)
I can reliably get a core dump with the short sequence ================== options { directory "/var/named" ; } ; include "db/conf.db"; ================== where /var/named/db/conf.db exists strace ended with: lstat("/etc/named.conf", {st_mode=S_IFREG|0664, st_size=1469, ...}) = 0 open("/etc/named.conf", O_RDONLY) = 19 fstat(19, {st_mode=S_IFREG|0664, st_size=1469, ...}) = 0 mmap(0, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x402e8000 read(19, "options {\n\tdirectory \"/var/named"..., 4096) = 1469 open("db/conf.db", O_RDONLY) = -1 ENOENT (No such file or directory) --- SIGSEGV (Segmentation fault) --- +++ killed by SIGSEGV +++ Hope this helps trace the problemI can reliably get a core dump with the short sequence ================== options { directory "/var/named" ; } ; include "db/conf.db"; ================== where /var/named/db/conf.db exists strace ended with: lstat("/etc/named.conf", {st_mode=S_IFREG|0664, st_size=1469, ...}) = 0 open("/etc/named.conf", O_RDONLY) = 19 fstat(19, {st_mode=S_IFREG|0664, st_size=1469, ...}) = 0 mmap(0, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x402e8000 read(19, "options {\n\tdirectory \"/var/named"..., 4096) = 1469 open("db/conf.db", O_RDONLY) = -1 ENOENT (No such file or directory) --- SIGSEGV (Segmentation fault) --- +++ killed by SIGSEGV +++ Hope this helps trace the problem
1.16r10-2 released, which is supposed to fix this bug.