Hi folks,
Have there been any changes to the way ntpd behaves in the past few updates? Reason I ask, whenever I run the ntpq or ntpdc, I immediately get a response of Name or service not known. I know things used to work sometime back, and never saw this kind of stuff. Nothing has changed in the /etc/ntp.conf file, nor has anything changed in the /etc/ntp/ntpservers files.
T hanks......
Sam
On 24/05/06, Sam Drinkard sam@wa4phy.net wrote:
Have there been any changes to the way ntpd behaves in the past few updates? Reason I ask, whenever I run the ntpq or ntpdc, I immediately get a response of Name or service not known. I know things used to work sometime back, and never saw this kind of stuff. Nothing has changed in the /etc/ntp.conf file, nor has anything changed in the /etc/ntp/ntpservers files.
According to the changelog of the NTP package nothing's changed since '04!
$ rpm -q --changelog ntp | less
"Name or service not known" sounds more like a resolution issue to me. Do you by any chance have caching-nameserver installed and had modified /etc/named.conf ? If so a caching-nameserver update may've overwritten your previous modifcations and potentially have broken DNS.
Are you using N.pool.ntp.org in /etc/ntp.conf? Can you resolve those hostnames?
$ awk '/ntp.org/ { print $NF }' /etc/ntp.conf | xargs -n1 dig +noall +answer
Will.
Will McDonald wrote:
On 24/05/06, Sam Drinkard sam@wa4phy.net wrote:
Have there been any changes to the way ntpd behaves in the past few updates? Reason I ask, whenever I run the ntpq or ntpdc, I immediately get a response of Name or service not known. I know things used to work sometime back, and never saw this kind of stuff. Nothing has changed in the /etc/ntp.conf file, nor has anything changed in the /etc/ntp/ntpservers files.
According to the changelog of the NTP package nothing's changed since '04!
$ rpm -q --changelog ntp | less
"Name or service not known" sounds more like a resolution issue to me. Do you by any chance have caching-nameserver installed and had modified /etc/named.conf ? If so a caching-nameserver update may've overwritten your previous modifcations and potentially have broken DNS.
Are you using N.pool.ntp.org in /etc/ntp.conf? Can you resolve those hostnames?
$ awk '/ntp.org/ { print $NF }' /etc/ntp.conf | xargs -n1 dig +noall +answer
Will.
Will,
Everything has worked correctly in the past, and I have not made any changes to the config files, nor dns, which is not a caching server. A little bit more info.. somehow, the machine on which I'm having this issue is i386, where the machine here at home is x86-64. It works as advertized. I copied all the config files over and deleted the first entry from the hosts, which on the local machine pointed to the remote machine. Other than that, they both are identical now. If, when executing either command, I enter "host vortex.wa4phy.net" then I get normal behavior. Somehow, it is not setting the default host, or a default host. I'm totally at a loss to explain unless there are some diffs between the two arch's.
Sam
Sam Drinkard wrote:
<snip>
Will,
Everything has worked correctly in the past, and I have not made any changes to the config files, nor dns, which is not a caching server. A little bit more info.. somehow, the machine on which I'm having this issue is i386, where the machine here at home is x86-64. It works as advertized. I copied all the config files over and deleted the first entry from the hosts, which on the local machine pointed to the remote machine. Other than that, they both are identical now. If, when executing either command, I enter "host vortex.wa4phy.net" then I get normal behavior. Somehow, it is not setting the default host, or a default host. I'm totally at a loss to explain unless there are some diffs between the two arch's.
Sam
Sam, just for grins, why don't you try replacing host names with IP addresses in your /etc/ntp.conf (after copying the original somewhere safe!), issue a #service ntpd restart and see what that does. Another thought: If you have a single-core processor accidentally running an smp kernel, ntpd will never stabilize. At least, that's the way it works on this AMD Athlon I'm using right now.
Robert wrote:
Sam Drinkard wrote:
<snip>
Will,
Everything has worked correctly in the past, and I have not made any changes to the config files, nor dns, which is not a caching server. A little bit more info.. somehow, the machine on which I'm having this issue is i386, where the machine here at home is x86-64. It works as advertized. I copied all the config files over and deleted the first entry from the hosts, which on the local machine pointed to the remote machine. Other than that, they both are identical now. If, when executing either command, I enter "host vortex.wa4phy.net" then I get normal behavior. Somehow, it is not setting the default host, or a default host. I'm totally at a loss to explain unless there are some diffs between the two arch's.
Sam
Sam, just for grins, why don't you try replacing host names with IP addresses in your /etc/ntp.conf (after copying the original somewhere safe!), issue a #service ntpd restart and see what that does. Another thought: If you have a single-core processor accidentally running an smp kernel, ntpd will never stabilize. At least, that's the way it works on this AMD Athlon I'm using right now.
Will,
I reverted back to the original configuration, and in the process, looked at the system-config-services. I'm seeing with the ntpd service stopped, but the subsys is locked. There was a pid file, but removed it. I can try ip's but they recommend using hostnames for most places due to the fact that IP's might change (a lot of the gov't servers are like that). I'm still dorking around trying to find what is exactly locked -- what subsys ? Anyhow, the machine does not run the smp kernel, but the machine here IS smp and runs the smp kernel. I'm just baffled why with identical config files, I still get the service not known message. Weird......
Sam
On Wed, 2006-05-24 at 13:26 -0400, Sam Drinkard wrote:
Robert wrote:
Sam Drinkard wrote:
<snip>
Will,
Everything has worked correctly in the past, and I have not made any changes to the config files, nor dns, which is not a caching server. A little bit more info.. somehow, the machine on which I'm having this issue is i386, where the machine here at home is x86-64. It works as advertized. I copied all the config files over and deleted the first entry from the hosts, which on the local machine pointed to the remote machine. Other than that, they both are identical now. If, when executing either command, I enter "host vortex.wa4phy.net" then I get normal behavior. Somehow, it is not setting the default host, or a default host. I'm totally at a loss to explain unless there are some diffs between the two arch's.
Sam
Sam, just for grins, why don't you try replacing host names with IP addresses in your /etc/ntp.conf (after copying the original somewhere safe!), issue a #service ntpd restart and see what that does. Another thought: If you have a single-core processor accidentally running an smp kernel, ntpd will never stabilize. At least, that's the way it works on this AMD Athlon I'm using right now.
Will,
I reverted back to the original configuration, and in the process,looked at the system-config-services. I'm seeing with the ntpd service stopped, but the subsys is locked.
Folks often forget this
ls /var/lock/subsys
should have an ntpd file.
There was a pid file, but removed it. I can try ip's but they recommend using hostnames for most places due to the fact that IP's might change (a lot of the gov't servers are like that). I'm still dorking around trying to find what is exactly locked -- what subsys ? Anyhow, the machine does not run the smp kernel, but the machine here IS smp and runs the smp kernel. I'm just baffled why with identical config files, I still get the service not known message. Weird......
Sam
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
William L. Maltby wrote:
On Wed, 2006-05-24 at 13:26 -0400, Sam Drinkard wrote:
Robert wrote:
Sam Drinkard wrote:
<snip>
Will,
Everything has worked correctly in the past, and I have not made any changes to the config files, nor dns, which is not a caching server. A little bit more info.. somehow, the machine on which I'm having this issue is i386, where the machine here at home is x86-64. It works as advertized. I copied all the config files over and deleted the first entry from the hosts, which on the local machine pointed to the remote machine. Other than that, they both are identical now. If, when executing either command, I enter "host vortex.wa4phy.net" then I get normal behavior. Somehow, it is not setting the default host, or a default host. I'm totally at a loss to explain unless there are some diffs between the two arch's.
Sam
Sam, just for grins, why don't you try replacing host names with IP addresses in your /etc/ntp.conf (after copying the original somewhere safe!), issue a #service ntpd restart and see what that does. Another thought: If you have a single-core processor accidentally running an smp kernel, ntpd will never stabilize. At least, that's the way it works on this AMD Athlon I'm using right now.
Will,
I reverted back to the original configuration, and in the process, looked at the system-config-services. I'm seeing with the ntpd service stopped, but the subsys is locked.
Folks often forget this
ls /var/lock/subsys
should have an ntpd file.
All is well on wa4phy.net. The whole thing boiled down to the corrupt entry in the /etc/hosts file for the loopback. It's a real wonder other stuff has not complained.....
Yes, I forgot about /var/lock/subsys too....
Many thanks.......
Sam
Robert wrote:
Sam Drinkard wrote:
<snip>
Will,
Everything has worked correctly in the past, and I have not made any changes to the config files, nor dns, which is not a caching server. A little bit more info.. somehow, the machine on which I'm having this issue is i386, where the machine here at home is x86-64. It works as advertized. I copied all the config files over and deleted the first entry from the hosts, which on the local machine pointed to the remote machine. Other than that, they both are identical now. If, when executing either command, I enter "host vortex.wa4phy.net" then I get normal behavior. Somehow, it is not setting the default host, or a default host. I'm totally at a loss to explain unless there are some diffs between the two arch's.
Sam
Sam, just for grins, why don't you try replacing host names with IP addresses in your /etc/ntp.conf (after copying the original somewhere safe!), issue a #service ntpd restart and see what that does. Another thought: If you have a single-core processor accidentally running an smp kernel, ntpd will never stabilize. At least, that's the way it works on this AMD Athlon I'm using right now.
I found one problem. Of all the dumb things to make a mistake on, and when, I have no idea. /etc/hosts file had an entry of 127.0.0.` instead of 127.0.0.1 -- that corrected the host or service unknown error, but now, when I try to issue any command from either ntpdc or ntpq, it times out. Sumtin still ain't right somewhere, but I'll find it eventually. Other than that, I think ntp is updating and workign as advertised.
Sam
-
Robert -
Sam Drinkard -
Will McDonald -
William L. Maltby