Bugzilla Bug 87902

Arches, please, mark net-dbs/bind{,-tools}-9.2.5 stable. 9.2.5 fixes all known
bugs to date in gentoo (refer to ChangeLogs for id's) and contains at least one
security fix since 9.2.2-r4.

stable on x86 and amd64

[ebuild  N    ] net-dns/bind-9.2.5  +berkdb +bind-mysql +caps +dlz +doc +idn
+ipv6 +ldap +mysql +odbc +postgres (-selinux) +ssl 0 kB

---->

Calculating dependencies ...done!
>>> emerge (1 of 1) net-dns/bind-9.2.5 to /
>>> md5 src_uri ;-) bind-9.2.5.tar.gz
>>> md5 src_uri ;-) dlz-9.2.5.patch.bz2
>>> Unpacking source...
>>> Unpacking bind-9.2.5.tar.gz to /var/tmp/portage/bind-9.2.5/work
>>> Unpacking dlz-9.2.5.patch.bz2 to /var/tmp/portage/bind-9.2.5/work
 * Applying dlz-9.2.5.patch.bz2 ...

 * Failed Patch: dlz-9.2.5.patch.bz2 !
 *  ( /var/tmp/portage/bind-9.2.5/temp/1667.patch )
 *
 * Include in your bugreport the contents of:
 *
 *   /var/tmp/portage/bind-9.2.5/temp/dlz-9.2.5.patch.bz2-1667.out

Gustavo: thanks, that's fixed now. stupid bug, indeed. :-(

sparc done now, thanks :)

bind-tools-9.2.5 stable on ppc.

btw, bind and bind-tools should be in sync.

bind stable on ppc.

This package was marked as stable by gmsoft. I remove hppa@g.o from the CC.

Is there any reason you've required >=mysql-4 in the use flags section? This
package is masked (package.mask) and therefore prevents the upgrade of BIND on
systems with the mysql use flag. Thanks.

My apologies. Outdated profile / mask file.

Is there, however, a specific reason to require MySQL 4? Are there any problems with 3, or any advantages to moving to 4 that would neccesitate such a change?

Stewart: mysql-3.23 is a bit outdated, imho. but, anyway, i'll check it.

I've some problems using bind (on ppc64) (never realy used it before..)

I've set up the chroot environment using "ebuild .. config". Now I edited /etc/conf.d/named: CPU="2". Then I try to start bind, but have some problems doing that:

G5 ~ # /etc/init.d/named start
 * Caching service dependencies ...                   [ ok ]
 * Starting chrooted named ...                        [ !! ]
G5 ~ # tail -n2 /var/log/messages 
Apr  5 15:35:10 G5 named[8692]: starting BIND 9.2.5 -u named -n 2 -t /chroot/dns
Apr  5 15:35:10 G5 named[8692]: using 1 CPU
G5 ~ # bind -u named -n 2 -t /chroot/dns
-bash: bind: -n: invalid option
bind: usage: bind [-lpvsPVS] [-m keymap] [-f filename] [-q name] [-u name] [-r keyseq] [-x keyseq:shell-command] [keyseq:readline-function or readline-command]
G5 ~ # 


So two problems:
- log shows "using 1 CPU" instead of 2
- bind doesn't use the -n option and fails to start

Any help for me? ;-)

Markus: i can't reproduce it on x86/amd64 boxes. please, recheck your settings.

shadanakar voxus # ebuild /var/db/pkg/net-dns/bind-9.2.5/bind-9.2.5.ebuild config
 *
 * Setting up the chroot directory... Done.
 *
 * Add the following to your root .bashrc or .bash_profile:
 *    alias rndc='rndc -k /chroot/dns/etc/bind/rndc.key'
 * Then do the following:
 *    source /root/.bashrc or .bash_profile
 *

shadanakar voxus # vi /etc/conf.d/named
shadanakar voxus # /etc/init.d/named start
 * Caching service dependencies ...                                                                                                    [ ok ]
 * Stopping chrooted named ...                                                                                                         [ ok ]
shadanakar voxus # tail -n 8 /var/log/messages
Apr  5 19:57:36 shadanakar named[15966]: starting BIND 9.2.5 -u named -n 2 -t /chroot/dns
Apr  5 19:57:36 shadanakar named[15966]: using 1 CPU
Apr  5 19:57:36 shadanakar named[15966]: loading configuration from '/etc/bind/named.conf'
Apr  5 19:57:36 shadanakar named[15966]: listening on IPv4 interface lo, 127.0.0.1#53
Apr  5 19:57:36 shadanakar named[15966]: command channel listening on 127.0.0.1#953
Apr  5 19:57:36 shadanakar named[15966]: zone 127.in-addr.arpa/IN: loaded serial 2002081601
Apr  5 19:57:36 shadanakar named[15966]: zone localhost/IN: loaded serial 2002081601
Apr  5 19:57:36 shadanakar named[15966]: running

but problem with only 1 processor usage is known - it's because of disabled threading in bind (grep thr bind-9.2.5.ebuild), because it's known to be buggy atm. i'm planning to enable it (via use flag) in next bump(s?) of 9.2.

Alpha done.

I just unmerged bind and deleted all configuration files, but that didn't
helped. I downgraded to latest stable version on ppc64 and all works fine.

I'll try to figure out, what is wrong with latest version, if I get some more
free time.

uh.. the first sentence should be: I just unmerged bind and deleted all
configuration files and reemerged bind, but that didn't helped.

ok, so 9.3 goes into testing now.
Markus: please, open separate bug, if that issue won't disappear. :-)

reopening for bind-tools stabilization

arm, ia64, s390, mips: please, mark stable bind-tools-9.2.5

Stable on mips.

arm/ia64/s390 done

so only ppc remains.

bind-9.2.5 stable on ppc.

hansmi: what's about bind-tools-9.2.5 on ppc64? that's the last one.

close it, if bind-tools is ok on ppc64, please.

and thanks to everyone. :-)

voxus: ppc != ppc64. I don't have a ppc64 installation by now, so I'll forward
this to the ppc64 team.

ok, reassigning

ppc64 is taken care of.

uff. fixed at last. :-)