Bind problems - FIXED
-
- Forum User
- Posts: 93
- Joined: Sun Feb 13, 2005 2:24 pm
- Location: TN
Bind problems - FIXED
CentOS 4, Plesk 8.4
Yesterday, I updated caching-nameserver and freetype. Today, all of the sudden, Bind just quit working. I tried to restart it, and got a message about rndc.key being missing or invalid, so I followed some instructions I found to regenerate the key and update named.conf.
Now, named starts fine, and the status shows that it's running with 37 zones loaded. However, whenever I do a lookup against it, I get SERVFAIL. I've verified that all the entries are in named.conf and restarted the service several times - I can't seem to pinpoint what the problem is.
Has anyone else had any trouble with the recent update of caching-nameserver (I'm assuming that was the culprit)? Any advice as to what to look at next?
UPDATE: I fixed it. Somehow, in named.conf, the directory option got set to /var/named instead of /var. Once I updated that and restarted, it worked.
Yesterday, I updated caching-nameserver and freetype. Today, all of the sudden, Bind just quit working. I tried to restart it, and got a message about rndc.key being missing or invalid, so I followed some instructions I found to regenerate the key and update named.conf.
Now, named starts fine, and the status shows that it's running with 37 zones loaded. However, whenever I do a lookup against it, I get SERVFAIL. I've verified that all the entries are in named.conf and restarted the service several times - I can't seem to pinpoint what the problem is.
Has anyone else had any trouble with the recent update of caching-nameserver (I'm assuming that was the culprit)? Any advice as to what to look at next?
UPDATE: I fixed it. Somehow, in named.conf, the directory option got set to /var/named instead of /var. Once I updated that and restarted, it worked.
-
- Forum User
- Posts: 93
- Joined: Sun Feb 13, 2005 2:24 pm
- Location: TN
Oh, sure - if you want to do it the easy way. I prefer to research vague error messages and run down rabbit holes for a couple of hours first...
What threw me off was that I updated it yesterday, and everything was fine. Then suddenly it quit working today. I now realize that it actually did quit working yesterday, but DNS caching prevented me from noticing until the records started expiring today.
Lesson learned.
What threw me off was that I updated it yesterday, and everything was fine. Then suddenly it quit working today. I now realize that it actually did quit working yesterday, but DNS caching prevented me from noticing until the records started expiring today.
Lesson learned.
-
- Long Time Forum Regular
- Posts: 2813
- Joined: Sat Aug 20, 2005 9:30 am
- Location: The Netherlands
The caching-nameserver package should not be installed on your Plesk server as it is known to cause problems (just like bind-chroot). See this knowledge base article: http://kb.parallels.com/en/234 (but don't exclude kernel updates like they tell you in that KB article).
Lemonbit Internet Dedicated Server Management
-
- Forum User
- Posts: 93
- Joined: Sun Feb 13, 2005 2:24 pm
- Location: TN
I just checked /etc/sysconfig/rhn/up2date and noticed that kernel* was in pkgSkipList. So removed it and ran up2date --dry-run, this was the result:
I take it that these updates should be added and won't cause any issues with Atomic on RHEL4/Plesk 8.4 running on a single Intel Core2 Duo processor?
Should /etc/yum.conf exclude sendmail bind-chroot caching-nameserver as suggested on the KB? It has entries as follows:
Thanks
Code: Select all
Testing package set / solving RPM inter-dependencies...
########################################
Name Version Rel
kernel-smp 2.6.9 67.0.20.EL i686
kernel-utils 2.4 13.1.105 i386
Should /etc/yum.conf exclude sendmail bind-chroot caching-nameserver as suggested on the KB? It has entries as follows:
Code: Select all
installonlypkgs=kernel kernel-smp kernel-devel kernel-smp-devel kernel-largesmp kernel-largesmp-devel kernel-hugemem kernel-hugemem-devel
-
- Long Time Forum Regular
- Posts: 2813
- Joined: Sat Aug 20, 2005 9:30 am
- Location: The Netherlands
This is the current RHEL kernel, so no, that shouldn't cause any problem with RHEL.Kalimari wrote:I just checked /etc/sysconfig/rhn/up2date and noticed that kernel* was in pkgSkipList. So removed it and ran up2date --dry-run, this was the result:I take it that these updates should be added and won't cause any issues with Atomic on RHEL4/Plesk 8.4 running on a single Intel Core2 Duo processor?Code: Select all
Testing package set / solving RPM inter-dependencies... ######################################## Name Version Rel kernel-smp 2.6.9 67.0.20.EL i686 kernel-utils 2.4 13.1.105 i386

As long as you don't install these packages you won't have any problems. You could explicitly exclude them to protect yourself from installing them, but you can also just not install them.Should /etc/yum.conf exclude sendmail bind-chroot caching-nameserver as suggested on the KB?

See the man page for yum.conf. The exclude and installonlypkgs directives are not the same (quite the opposite actually). Do you use both up2date and yum?It has entries as follows:Code: Select all
installonlypkgs=kernel kernel-smp kernel-devel kernel-smp-devel kernel-largesmp kernel-largesmp-devel kernel-hugemem kernel-hugemem-devel
Lemonbit Internet Dedicated Server Management
-
- Long Time Forum Regular
- Posts: 2813
- Joined: Sat Aug 20, 2005 9:30 am
- Location: The Netherlands
If you don't mind your OS slowly changing into CentOS (which doesn't come with paid support, but is 100% compatible with RHEL otherwise) you could move everything to yum, which can also install updates automatically if you like, and deal with all updates through one program.
Lemonbit Internet Dedicated Server Management