Not sure what broke Mara...

Greg Platt - Platt Consultants GregPlatt at ix.netcom.com
Mon Aug 18 02:30:19 EDT 2008


Early this afternoon I finished a reinstall of Maradns on my production
server. Using my notes from my first test server build, I'm now in the
process of building my production server from scratch. I'm running the
latest Debian release on a stable etch 4.0r3 system and brand new server
hardware.

 

I completed the install of the second IP address and tested to confirm it
was activated. Then I recreated the maradns' config files and tested to
confirm they loaded with no problems. Once they were working fine, I started
the server and confirmed that maradns was running. Everything looked
perfect. I was able to reach both of the 2 production sites I'm testing with
without problems.

 

My next step was to use the test sites to get my CGI scripts working so,
having finished with mara (or so I naively thought!) I went on to work on
getting the CGI scripts to run in the two test sites as well. I got the CGI
process working without problems. And I rebooted the server to confirm that
all my changes still worked.

 

I was shocked to find that one of my 2 test sites came up fine but the other
did not. IN fact, I can't even PING the site. It's as if mara was completely
refusing to reply to requests for it. To try to figure out what had
happened, I restarted mara manually again and discovered that it is now
throwing "fatal errors on line 26" message for all 3 sites my name server
site and the 2 test sites). When running in stand alone mode. This worked
this afternoon. I KNOW it did. I didn't even TRY to run mara in resident
mode until it had started without errors in stand alone mode. 

 

Here's what I see from the ifconfig. 

 

[code]eth1      Link encap:Ethernet  HWaddr 00:1F:C6:76:C1:A2

          inet addr:208.19.190.68  Bcast:208.19.191.255  Mask:255.255.254.0

          inet6 addr: fe80::21f:c6ff:fe76:c1a2/64 Scope:Link

          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

          RX packets:1411 errors:0 dropped:0 overruns:0 frame:0

          TX packets:0 errors:0 dropped:0 overruns:0 carrier:1

          collisions:0 txqueuelen:1000

          RX bytes:145501 (142.0 KiB)  TX bytes:0 (0.0 b)

          Memory:febc0000-fec00000

 

eth1:0    Link encap:Ethernet  HWaddr 00:1F:C6:76:C1:A2

          inet addr:208.19.190.69  Bcast:208.19.191.255  Mask:255.255.254.0

          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

          Memory:febc0000-fec00000

 

lo        Link encap:Local Loopback

          inet addr:127.0.0.1  Mask:255.0.0.0

          inet6 addr: ::1/128 Scope:Host

          UP LOOPBACK RUNNING  MTU:16436  Metric:1

          RX packets:13 errors:0 dropped:0 overruns:0 frame:0

          TX packets:13 errors:0 dropped:0 overruns:0 carrier:0

          collisions:0 txqueuelen:0

          RX bytes:975 (975.0 b)  TX bytes:975 (975.0 b)[/code]

It tells me both IP addresses are active.

 

Here's what I see when I reboot the server. Maradns seems to be running in
resident mode

 

[code]myserver:~# ps aux |grep mara

nobody    2342  0.0  0.0   3944   724 ?        S    23:04   0:00
/usr/sbin/maradns -f /etc/maradns/mararc

root      2343  0.0  0.0   2640   532 ?        S    23:04   0:00 logger -p
daemon.notice -t maradns.etc_maradns_mararc

root      2763  0.0  0.0   3936   692 pts/0    D+   23:06   0:00 grep
mara[/code]

But HERE is what I see when I run mara from the command prompt

 

[code]myserver:/etc/maradns# maradns

Using default ICANN root servers

 Log: Root directory changed

 Log: Binding to address 208.19.190.68, 208.19.190.69

 Log: Socket opened on UDP port 53

 Log: Root privileges dropped

Warning: Fatal error in zone file testsite2.com. (aborting this zone file)
on line 26

Warning: Fatal error in zone file testsite1.com. (aborting this zone file)
on line 26

Warning: Fatal error in zone file myhost.com. (aborting this zone file) on
line 26

 Log: All RRs have been loaded[/code]

 

These errors were NOT showing up this afternoon. I swear they were not. I
didn't even TRY to put mara 

into resident mode until all the errors at this level had been eliminated.

 

Here is one of the zone files involved this is the site that DOES still work
despite the "fatal error" reported 

# This is an example zone file for the imaginary domain example.com.

 

# To adapt this file to another zone, we need to make the following changes:

# Substitute the name example.com with the name of the domain one wishes to

# administer.

# Substitute the IP 10.10.10.12 with the IP of the machine running the web

# server for the domain.

# Substitute the IP 10.10.10.15 with the IP of the machine running the mail

# server for the domain.

# Note that the chroot_dir mararc variable tells MaraDNS which directory

# contains domain zone files.

 

# example.com.          NS      ns1.example.com.

# ns1.example.com.              10.10.10.19

# example.com.                  10.10.10.19

# www.example.com.              10.10.10.19

# example.com.          MX 10   mail1.example.com.

# mail1.example.com.            10.10.10.19

 

# First we must define the name server for the domain and the name server's

# address

 

# example.com.          NS      ns1.example.com.

# ns1.example.com.              10.10.10.19

 

testsite2.com.       NS      ns1.myhost.com. <--- Note: this is LINE 26!

testsite2.com.       NS      ns2.myhost.com.

ns1.myhost.com.                208.19.190.68

ns2.myhost.com.                208.19.190.69

 

# We have two records which handle the serving of web pages.  These

# are simple name-to-ip translations. In other words, we have one

# record which states "The IP for example.com. is 10.10.10.12" and

# another that states "The IP for www.example.com. is 10.10.10.12"

 

# example.com.                  10.10.10.12

# www.example.com.              10.10.10.12

 

testsite2.com.               208.19.190.68

www.testsite2.com.           208.19.190.68

testsite2.com.               208.19.190.69

www.testsite2.com.           208.19.190.69

 

# We also need to set up an IP for the machine that serves email for

# example.com.  Because of how DNS is designed, we both need[/code]

 

Yet, DESPITE the reported fatal errors the test site referenced in THIS file
comes up and displays

on the web just fine. While the other test site which I ran and tested with
for 3 full hours now 

fails to respond to ping and does not come up on the web at all. delivering
what amounts to a 

404 error everytime I try to access it.

 

So what the heck is WRONG here? Can somebody give me a clue please? I'm
totally baffled. And at 

the moment it's after midnight so I'm gonna go crash and hope that mara has
healed itself in the 

morning.

 

Thanks!



More information about the list mailing list