Re: dhcpd (reprise)

2004-11-09 Thread Lowell Gilbert
Gary Kline [EMAIL PROTECTED] writes:

 On Mon, Nov 08, 2004 at 09:29:47AM -0500, Lowell Gilbert wrote:
  Gary Kline [EMAIL PROTECTED] writes:
  
   On Sun, Nov 07, 2004 at 08:56:58PM -0500, Lowell Gilbert wrote:
   
   
   No subnet declaration for dc0 (216.231.43.140).
   ** Ignoring requests on dc0.  If this is not what
  you want, please write a subnet declaration
  in your dhcpd.conf file for the network segment
  to which interface dc0 is attached. **
   
   Sending on   Socket/fallback/fallback-net
   
 I've seen this before.  What does the last line mean?
 Or, how do I test this?  I've just tried ssh'ing
 around.  Nothing to the screen.
  
  
  This is telling you that the machine doesn't know how to assign
  addresses for DHCP requests that come in on the dc0 interface.  If
  that's correct (i.e., you want it to assign addresses on some other
  interface but not that one), then everything's fine so far.  If it's
  not, then you need to modify your dhcpd.conf as it said.
 
   I've got two NICs on my primary.  dc0 goes to my router;
   dc1 goes to my hub.  All are running unix.  So far, I 
   have rebooted only my laptop.  I can immediately ssh from
   my latop *into* my primary (DNS) server, but when I try to
   ping anywhere from my laptop, nothing--it times out.
   So my dhcpd isn't handing out leases.  

Eh?  If you can ssh into the DNS server, you have an address.

   In /etc/rc.conf I've got:
 
   dhcpd_flags=-q# command option(s)
   dhcpd_conf=/usr/local/etc/dhcpd.conf# configuration file
   dhcpd_ifaces=dc1  # ethernet interface(s)
   dhcpd_withumask=022   # file creation mask
 
   So far, the dhcpd_ifaces doesn't seem to be working, 
   although I *do* see it when I do a grep on 
   'sh -x on /usr/local/etc/rc.d/isc-dhcpd.sh::
 
   + network_interfaces=dc0 dc1 lo0
   + ifconfig_dc1=inet 10.0.0.1  netmask 255.0.0.0
   + dhcpd_ifaces=dc1
 
   So the script at least is reading /etc/rc.conf.  Why dhcpd
   isn't seeing this is unknown.
 
 
   Here is part of my dhcpd.conf:
 
 
 option dhcp-server-identifier 10.0.0.1;
 option domain-name thought.org;
 option domain-name-servers 216.231.41.2, 66.93.87.2;
 option routers 10.0.0.1;
 option subnet-mask 255.0.0.0;
 server-name sage;
 server-identifier 10.0.0.1;

 
  
  If everything is okay on that front, then you need to get some of the
  other machines (the ones to which this server should be assigning
  addresses) to ask for leases.  How to do this depends on what OS they
  are running, but rebooting should do it in any case.
 
 
   So far, rebooting ns1.thought.org (== sage) and my laptop
   don't change anything.

It *did* force the laptop to try to renegotiate its lease.  
Verbose dhcpd output from ns1 at that time would have told you what
the problem was if it in fact was related to DHCP.

It sounds, though, as though the real problem is related to DNS.
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-09 Thread Gary Kline
On Tue, Nov 09, 2004 at 11:42:54AM -0500, Lowell Gilbert wrote:
 Gary Kline [EMAIL PROTECTED] writes:
 
  On Mon, Nov 08, 2004 at 09:29:47AM -0500, Lowell Gilbert wrote:
   Gary Kline [EMAIL PROTECTED] writes:


[[ Last night mail suddenly quit working.  There is
nothng/zero in /var/log/maillog, so I'm checking with
my ISP.  Meanwhile, writing from my alt site 
while I still can.  ... ]]
   
 Eh?  If you can ssh into the DNS server, you have an address.


My DNS box has my only real IP, 216.231.43.140.  I can
ssh *into* sage/ns1 from my laptops at once; but sshing 
(*out*) takes at least 2 minutes.  This is a major mystery
except that in dhcpd.conf I have most of my private 
servers 'hardwired';  the one that is not listed is not
reachable.

 
  
  So far, rebooting ns1.thought.org (== sage) and my laptop
  don't change anything.
 
 It *did* force the laptop to try to renegotiate its lease.  
 Verbose dhcpd output from ns1 at that time would have told you what
 the problem was if it in fact was related to DHCP.

Hmmm. Is this the attempt? (10.249) is the laptop.  

Nov  8 21:10:45 sage dhcpd: DHCPDISCOVER from 00:10:a4:06:cb:39 via dc1
Nov  8 21:10:45 sage dhcpd: DHCPOFFER on 10.0.0.249 to 00:10:a4:06:cb:39 via dc1
Nov  8 21:10:45 sage dhcpd: Dynamic and static leases present for 10.0.0.249.
Nov  8 21:10:45 sage dhcpd: Remove host declaration zen or remove 10.0.0.249
Nov  8 21:10:45 sage dhcpd: from the dynamic address pool for 10.0/8
Nov  8 21:10:45 sage dhcpd: DHCPREQUEST for 10.0.0.249 (10.0.0.1) from 
00:10:a4:06:cb:39 via dc1
Nov  8 21:10:45 sage dhcpd: DHCPACK on 10.0.0.249 to 00:10:a4:06:cb:39 via dc1
Nov  9 00:10:12 sage dhcpd: Dynamic and static leases present for 10.0.0.247.
Nov  9 00:10:12 sage dhcpd: Remove host declaration tao or remove 10.0.0.247
Nov  9 00:10:12 sage dhcpd: from the dynamic address pool for 10.0/8
Nov  9 00:10:12 sage dhcpd: DHCPREQUEST for 10.0.0.247 from 00:d0:b7:f0:de:ea 
via dc1
Nov  9 00:10:12 sage dhcpd: DHCPACK on 10.0.0.247 to 00:d0:b7:f0:de:ea via dc1

(Again, sshing from 10.249 to 10.1 is immediate; sshing from 10.1 out
to, say, 10.247 takes a couple minutes...)  Also, if the daemon is 
issuing
a lease, why can't I ping from the laptop? or any other server?


 
 It sounds, though, as though the real problem is related to DNS.


You may be right, altho www.dnsreport.com doesn't see anything.
If I get my DNS files to you, can you tell?  --I've been using
bind-9 going on 4 years.  After kiddie scripts kept bumping bind-8
off.  I'm using v9.3,0 and maybe somethng is misconfigured.

gary


-- 
Gary Kline  Seattle BSD Users' Group (seabug)  | [EMAIL PROTECTED]
Thought Unlimited Org's Alternate Email Site
http://www.magnesium.net/~kline
   To live is not a necessity; but to live honorably...is a necessity. -Kant

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-08 Thread Lowell Gilbert
Gary Kline [EMAIL PROTECTED] writes:

 On Sun, Nov 07, 2004 at 08:56:58PM -0500, Lowell Gilbert wrote:
  Gary Kline [EMAIL PROTECTED] writes:
  
 I'll add the -d flag to my rc.conf; are you using the latest
  
  No; *disable* dhcpd from rc.conf, and start it by hand with the -d flag.
  
 version of the dhcp-server from ISC or something other?
  
  I'm running the ISC dhcp server, as I mentioned in my message, and
  it's fairly up-to-date, but I don't think the exact version matters
  for you (at least not at this point).
  
 I looked for a 'debug' flag on the man page and didn't see it.
  
  From man dhcpd:
 To have dhcpd log to the standard  error  descriptor,  specify  the  
  -d
 flag.  This can be useful for debugging, and also at sites where a 
  com-
 plete log of all dhcp activity must be kept but syslogd is not 
  reliable
 or  otherwise  cannot  be  used.Normally, dhcpd will log all 
  output
 using the syslog(3) function with the log facility set to LOG_DAEMON.
  
 What should I loook for in th logfile?  or will it be 
 obvious :-)
  
  Again, the approach I'm describing will *not* log into the logfile;
  I'm suggesting you get the debug output on a console in real time.
 
 
   Okay. This is all that is output to stderr:
 
 
 No subnet declaration for dc0 (216.231.43.140).
 ** Ignoring requests on dc0.  If this is not what
you want, please write a subnet declaration
in your dhcpd.conf file for the network segment
to which interface dc0 is attached. **
 
 Sending on   Socket/fallback/fallback-net
 
   I've seen this before.  What does the last line mean?
   Or, how do I test this?  I've just tried ssh'ing
   around.  Nothing to the screen.


This is telling you that the machine doesn't know how to assign
addresses for DHCP requests that come in on the dc0 interface.  If
that's correct (i.e., you want it to assign addresses on some other
interface but not that one), then everything's fine so far.  If it's
not, then you need to modify your dhcpd.conf as it said.

If everything is okay on that front, then you need to get some of the
other machines (the ones to which this server should be assigning
addresses) to ask for leases.  How to do this depends on what OS they
are running, but rebooting should do it in any case.
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-08 Thread Gary Kline
On Mon, Nov 08, 2004 at 09:29:47AM -0500, Lowell Gilbert wrote:
 Gary Kline [EMAIL PROTECTED] writes:
 
  On Sun, Nov 07, 2004 at 08:56:58PM -0500, Lowell Gilbert wrote:
  
  
  No subnet declaration for dc0 (216.231.43.140).
  ** Ignoring requests on dc0.  If this is not what
 you want, please write a subnet declaration
 in your dhcpd.conf file for the network segment
 to which interface dc0 is attached. **
  
  Sending on   Socket/fallback/fallback-net
  
  I've seen this before.  What does the last line mean?
  Or, how do I test this?  I've just tried ssh'ing
  around.  Nothing to the screen.
 
 
 This is telling you that the machine doesn't know how to assign
 addresses for DHCP requests that come in on the dc0 interface.  If
 that's correct (i.e., you want it to assign addresses on some other
 interface but not that one), then everything's fine so far.  If it's
 not, then you need to modify your dhcpd.conf as it said.

I've got two NICs on my primary.  dc0 goes to my router;
dc1 goes to my hub.  All are running unix.  So far, I 
have rebooted only my laptop.  I can immediately ssh from
my latop *into* my primary (DNS) server, but when I try to
ping anywhere from my laptop, nothing--it times out.
So my dhcpd isn't handing out leases.  

In /etc/rc.conf I've got:

dhcpd_flags=-q# command option(s)
dhcpd_conf=/usr/local/etc/dhcpd.conf# configuration file
dhcpd_ifaces=dc1  # ethernet interface(s)
dhcpd_withumask=022   # file creation mask

So far, the dhcpd_ifaces doesn't seem to be working, 
although I *do* see it when I do a grep on 
'sh -x on /usr/local/etc/rc.d/isc-dhcpd.sh::

+ network_interfaces=dc0 dc1 lo0
+ ifconfig_dc1=inet 10.0.0.1  netmask 255.0.0.0
+ dhcpd_ifaces=dc1

So the script at least is reading /etc/rc.conf.  Why dhcpd
isn't seeing this is unknown.


Here is part of my dhcpd.conf:


option dhcp-server-identifier 10.0.0.1;
option domain-name thought.org;
option domain-name-servers 216.231.41.2, 66.93.87.2;
option routers 10.0.0.1;
option subnet-mask 255.0.0.0;
server-name sage;
server-identifier 10.0.0.1;

 
 If everything is okay on that front, then you need to get some of the
 other machines (the ones to which this server should be assigning
 addresses) to ask for leases.  How to do this depends on what OS they
 are running, but rebooting should do it in any case.


So far, rebooting ns1.thought.org (== sage) and my laptop
don't change anything.

gary


-- 
   Gary Kline [EMAIL PROTECTED]   www.thought.org Public service Unix

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-07 Thread Lowell Gilbert
Gary Kline [EMAIL PROTECTED] writes:

   I don't know if my mail from last night (Friday, localtime)
   got thru or not.  Part of sendmail is bolloxed too...  I
   see my /var/log/maillog filling up, but unable to resolve
   mail from freebsd.org.  Or anywhere.
 
   Anyhow, last night thngs on my primary server broke.  I can
   ping, I can use mozilla, wshatever, but only from 
   ns1.thought.org.  
 
   I fixed the new bind9 paths and re-exec'd those.  But for
   unknown reasons, dhcpd fails to hand out new leases.  Does
   anybody have any ideas howto debug this?  (I've tried some
   people here on the Seattle list, but they're unavailable.)
 
   So: nutshell, looks like my /etc/namedb/* stuff is okay.
   --Something wrong with sendmail.--  And dhcp* quit working.
   Any real, hardcore system admins out there who can help me??

If sendmail is still failing to resolve names, then it sounds like
your named setup is *not* okay.  But you'd have to show us what the
symptoms are precisely in order for us to help figure it out.

For dhcpd, the answers are either in the dhcpd logs or could be if you
raised the verbosity.  I've been using the ISC server lately in a
development testbed, and I use the '-d' option to get the log
information sent directly to my terminal.  
 dhcpd -d -cf /usr/local/etc/dhcpd.debug.conf
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-07 Thread Gary Kline
On Sun, Nov 07, 2004 at 11:20:29AM -0500, Lowell Gilbert wrote:
 Gary Kline [EMAIL PROTECTED] writes:
 
  I don't know if my mail from last night (Friday, localtime)
  got thru or not.  Part of sendmail is bolloxed too...  I
  see my /var/log/maillog filling up, but unable to resolve
  mail from freebsd.org.  Or anywhere.
  
  Anyhow, last night thngs on my primary server broke.  I can
  ping, I can use mozilla, wshatever, but only from 
  ns1.thought.org.  
  
  I fixed the new bind9 paths and re-exec'd those.  But for
  unknown reasons, dhcpd fails to hand out new leases.  Does
  anybody have any ideas howto debug this?  (I've tried some
  people here on the Seattle list, but they're unavailable.)
  
  So: nutshell, looks like my /etc/namedb/* stuff is okay.
  --Something wrong with sendmail.--  And dhcp* quit working.
  Any real, hardcore system admins out there who can help me??
 
 If sendmail is still failing to resolve names, then it sounds like
 your named setup is *not* okay.  But you'd have to show us what the
 symptoms are precisely in order for us to help figure it out.
 
 For dhcpd, the answers are either in the dhcpd logs or could be if you
 raised the verbosity.  I've been using the ISC server lately in a
 development testbed, and I use the '-d' option to get the log
 information sent directly to my terminal.  
  dhcpd -d -cf /usr/local/etc/dhcpd.debug.conf

The sendmail problem that were printed to stderr went away
after I created a new 5.3 sendmail.cf; with dhcpd problems,
mail can't get to my private servers so I have it re-routed
to ns1.thought.org.

I'll add the -d flag to my rc.conf; are you using the latest
version of the dhcp-server from ISC or something other?
I looked for a 'debug' flag on the man page and didn't see it.
What should I loook for in th logfile?  or will it be 
obvious :-)

thank,

gary



-- 
Gary Kline  Seattle BSD Users' Group (seabug)  | [EMAIL PROTECTED]
Thought Unlimited Org's Alternate Email Site
http://www.magnesium.net/~kline
   To live is not a necessity; but to live honorably...is a necessity. -Kant

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-07 Thread Lowell Gilbert
Gary Kline [EMAIL PROTECTED] writes:

   I'll add the -d flag to my rc.conf; are you using the latest

No; *disable* dhcpd from rc.conf, and start it by hand with the -d flag.

   version of the dhcp-server from ISC or something other?

I'm running the ISC dhcp server, as I mentioned in my message, and
it's fairly up-to-date, but I don't think the exact version matters
for you (at least not at this point).

   I looked for a 'debug' flag on the man page and didn't see it.

From man dhcpd:
   To have dhcpd log to the standard  error  descriptor,  specify  the  -d
   flag.  This can be useful for debugging, and also at sites where a com-
   plete log of all dhcp activity must be kept but syslogd is not reliable
   or  otherwise  cannot  be  used.Normally, dhcpd will log all output
   using the syslog(3) function with the log facility set to LOG_DAEMON.

   What should I loook for in th logfile?  or will it be 
   obvious :-)

Again, the approach I'm describing will *not* log into the logfile;
I'm suggesting you get the debug output on a console in real time.
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: dhcpd (reprise)

2004-11-07 Thread Gary Kline
On Sun, Nov 07, 2004 at 08:56:58PM -0500, Lowell Gilbert wrote:
 Gary Kline [EMAIL PROTECTED] writes:
 
  I'll add the -d flag to my rc.conf; are you using the latest
 
 No; *disable* dhcpd from rc.conf, and start it by hand with the -d flag.
 
  version of the dhcp-server from ISC or something other?
 
 I'm running the ISC dhcp server, as I mentioned in my message, and
 it's fairly up-to-date, but I don't think the exact version matters
 for you (at least not at this point).
 
  I looked for a 'debug' flag on the man page and didn't see it.
 
 From man dhcpd:
To have dhcpd log to the standard  error  descriptor,  specify  the  -d
flag.  This can be useful for debugging, and also at sites where a com-
plete log of all dhcp activity must be kept but syslogd is not reliable
or  otherwise  cannot  be  used.Normally, dhcpd will log all output
using the syslog(3) function with the log facility set to LOG_DAEMON.
 
  What should I loook for in th logfile?  or will it be 
  obvious :-)
 
 Again, the approach I'm describing will *not* log into the logfile;
 I'm suggesting you get the debug output on a console in real time.


Okay. This is all that is output to stderr:


No subnet declaration for dc0 (216.231.43.140).
** Ignoring requests on dc0.  If this is not what
   you want, please write a subnet declaration
   in your dhcpd.conf file for the network segment
   to which interface dc0 is attached. **

Sending on   Socket/fallback/fallback-net

I've seen this before.  What does the last line mean?
Or, how do I test this?  I've just tried ssh'ing
around.  Nothing to the screen.

gary




-- 
Gary Kline  Seattle BSD Users' Group (seabug)  | [EMAIL PROTECTED]
Thought Unlimited Org's Alternate Email Site
http://www.magnesium.net/~kline
   To live is not a necessity; but to live honorably...is a necessity. -Kant

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


dhcpd (reprise)

2004-11-05 Thread Gary Kline
Guys,

I don't know if my mail from last night (Friday, localtime)
got thru or not.  Part of sendmail is bolloxed too...  I
see my /var/log/maillog filling up, but unable to resolve
mail from freebsd.org.  Or anywhere.

Anyhow, last night thngs on my primary server broke.  I can
ping, I can use mozilla, wshatever, but only from 
ns1.thought.org.  

I fixed the new bind9 paths and re-exec'd those.  But for
unknown reasons, dhcpd fails to hand out new leases.  Does
anybody have any ideas howto debug this?  (I've tried some
people here on the Seattle list, but they're unavailable.)

So: nutshell, looks like my /etc/namedb/* stuff is okay.
--Something wrong with sendmail.--  And dhcp* quit working.
Any real, hardcore system admins out there who can help me??

thanks much,

gary



-- 
Gary Kline  Seattle BSD Users' Group (seabug)  | [EMAIL PROTECTED]
Thought Unlimited Org's Alternate Email Site
http://www.magnesium.net/~kline
   To live is not a necessity; but to live honorably...is a necessity. -Kant

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]