Okay, so I need a new subnet at home - I want to use mrx for testing packet loss. I added another ethernet card to it and configured up the 192.168.3.0/24 subnet. So far it has two hosts: drx (2nd interface on mrx) and kanigix (second interface on sandman). I just plugged the two together and only configured drx:

dev8278   Link encap:Ethernet  HWaddr 00:03:47:AF:35:21
          inet addr:192.168.3.1  Bcast:192.168.3.255  Mask:255.255.255.0
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)

I didn't add any routing info or anything else. Okay, I tried pinging drx and it worked. Ssh? Failed. Why? Am I getting to the right place?

                                                  My traceroute  [v0.69]
mrx.internal.excfb.com (0.0.0.0)(tos=0x0 psize=64 bitpattern=0x00)                               Thu Mar 23 14:05:36 2006
Keys:  Help   Display mode   Restart statistics   Order of fields   quit
                                                                                        Packets               Pings
 Host                                                                                 Loss%  Last   Avg  Best  Wrst StDev
 1. janus.internal.excfb.com                                                           0.0%   0.6   0.6   0.6   0.7   0.1
 2. ip68-0-80-1.tu.ok.cox.net                                                          0.0%  28.5  16.2  10.2  28.5   8.3
 3. 68.12.9.249                                                                        0.0%  22.5  14.8  11.4  22.5   5.2
 4. drx.internal.excfb.com                                                             0.0%  19.1  24.2  17.7  38.9   9.9

Ugh, that scares me. I'm going out to my ISP and then somewhere. I'm not coming back into my network.

First off, I want to blog about it. I can't get to Kool Aid Served Daily. I can't get to google. Okay, I kill the second interface on wont, I unplug it, I do everything (alway using '/etc/init.d/network restart' to clear things). I still can't get out. I had to reconfigure the interface to use DHCP and keep it unplugged.

Back to the real problems:

• I need to tell mrx how to route traffic.
• I need to tell my firewall how to not pass out to private address space.

Great the Belkin isn't smart enough to block outgoing address spaces.

I'll have another entry about the Linux router, I think I've got needs just a bit different than most. Not complex, just different and hard to find good examples on google.

You know you've been doing too much network configuration when you pick up the phone and start off with 192 and wonder where the '.' is. This almost happened to me - Tulsa is area code 918, so I got to '19' before I realized I was down the wrong path.

With Fedora Core 4 or 5, I can mount via NFSv4 to a Solaris Nevada b35 system, but I can't mount the Fedora box from Solaris. The problem is I think known, i.e., we hit it at Connectathon 2006, and it affects all OSes which implement their NFSv4 stack via the Linux reference implementation.

Here is a mount from mrx (Stock Fedora Core 5) to wont (Stock Solaris b35):

[root@mrx wont]# uname -a
Linux mrx.internal.excfb.com 2.6.15-1.2054_FC5 #1 Tue Mar 14 15:48:33 EST 2006 i686 i686 i386 GNU/Linux
[root@mrx etc]# showmount -e wont
Export list for wont:
/export/zfs              (everyone)
/export/zfs/nfsv2        (everyone)
/export/zfs/nfsv3        (everyone)
/export/zfs/nfsv4        (everyone)
/export/zfs/tdh          (everyone)
/zoo/isos/mnt/freebsd/d2 (everyone)
/zoo/isos/mnt/Fedora/dvd (everyone)
[root@mrx etc]# mount -t nfs4 wont:/ /nfsv4/wont
[root@mrx etc]# cd /nfsv4/wont
[root@mrx wont]# ls -la
total 8
drwxr-xr-x 42 root root 1024 Mar 23 11:21 .
drwxr-xr-x  3 root root 4096 Mar 23 11:23 ..
drwxr-xr-x  4 root sys   512 Mar 21 15:28 export
drwxr-xr-x  4 root sys     4 Mar 20 23:14 zoo

And here is the corresponding attempt from wont to mrx:

[tdh@wont ~]> uname -a
SunOS wont 5.11 snv_35 i86pc i386 i86pc
[tdh@wont ~]> sudo mount mrx:/ /nfsv4/mrx
nfs mount: mount: /nfsv4/mrx: Resource temporarily unavailable

We need to make sure that NFSv4 is working correctly on mrx, so we try the same thing from adept (Heavily updated Fedora Core 4 system):

[tdh@adept ~]> uname -a
Linux adept.internal.excfb.com 2.6.15-1.1833_FC4 #1 Wed Mar 1 23:41:37 EST 2006 i686 i686 i386 GNU/Linux
[root@adept ~]# ls -la /nfsv4/mrx
total 204
drwxr-xr-x  24 root nobody  4096 Mar 23 11:23 .
drwxr-xr-x   4 root root    4096 Mar 23 11:39 ..
-rw-r--r--   1 root nobody     0 Mar 23 10:55 .autofsck
-rw-r--r--   1 root nobody     0 Mar 23 00:02 .autorelabel
drwxr-xr-x   2 root nobody  4096 Mar 23 02:06 bin
drwxr-xr-x   2 root nobody  4096 Mar 22 16:27 boot

Okay, first off, ID mapping is not working correctly here. Notice that the user is correct, but the group is not. Is 'root' on both systems?

[root@mrx wont]# grep root /etc/group
root:x:0:root
[root@adept ~]# grep root /etc/group
root:x:0:root

Yes, this should display correctly.

So, what is going on here with respect to b35 not being able to mount either box (not shown here is the attempt to mount off of adept, which actually started all of this off)?

We need to look at what is going across the wire.

[tdh@wont ~]> sudo snoop -x0,2000 -o w2m.snoop wont mrx
Using device /dev/bge0 (promiscuous mode)
60 ^C

So lets walk through some of the output. wont asks mrx if it speaks NFSv4 and the answer is yes.

 13   0.00018         wont -> mrx.internal.excfb.com NFS C NULL4
 15   0.00315 mrx.internal.excfb.com -> wont         NFS R NULL4

Okay, great, tell me what your root filehandle is...

 23   0.00005         wont -> mrx.internal.excfb.com NFS C 4 (mount       ) PUTROOTFH GETFH GETATTR c8000167 0
 25   0.00012 mrx.internal.excfb.com -> wont         NFS R 4 (mount       ) NFS4_OK PUTROOTFH NFS4_OK GETFH NFS4_OK FH=0015 GETATTR NFS4_OK

Sweet, hey, I have to register myself with you...

 27   0.00013         wont -> mrx.internal.excfb.com NFS C 4 (setclientid ) PUTROOTFH GETATTR 400 0 SETCLIENTID Prog=1073741824 ID=tcp Addr=192.168.2.105.186.185 CBID=107374...
 28   0.00020 mrx.internal.excfb.com -> wont         NFS R 4 (setclientid ) NFS4_OK PUTROOTFH NFS4_OK GETATTR NFS4_OK SETCLIENTID NFS4_OK CL=6cd9224406000000 CFV=B8DC2244543F1101

Hey, I trust you and all, but can we verify I heard you right?

 30   0.00003         wont -> mrx.internal.excfb.com NFS C 4 (sclntid_conf) SETCLIENTID_CONFIRM CL=6cd9224406000000 CFV=B8DC2244543F1101
 31   0.00013 mrx.internal.excfb.com -> wont         NFS R 4 (sclntid_conf) NFS4ERR_CLID_INUSE SETCLIENTID_CONFIRM NFS4ERR_CLID_INUSE

What do you mean, all of us Solaris boxes look alike to you?

The issue here is that the mrx gives a client identifier to wont and then says it is already assigned to another client.

A 'yum update' does not show a new kernel ready. But I do know this was looked at during Connectathon.

I sometimes need other OSes to do interoperability testing (and to play games). I had to double up my WinXP box (i.e., the one for games) with Linux. So I got the new Fedora Core 5 bits. It looks pretty.

One nit is that they have removed the 'Open Terminal' option from the right-click-on-desktop menu. Very annoying.

I also made a tarball of my home directories on my Fedora Core 4 system. I need the new system to do some packet routing and also to hopefully bypass the SETCLIENTID bug all of the Linux systems were showing at Connectathon 2006.

adept:

[tdh@adept ~]> uname -a
Linux adept.internal.excfb.com 2.6.15-1.1833_FC4 #1 Wed Mar 1 23:41:37 EST 2006 i686 i686 i386 GNU/Linux

mrx:

[tdh@mrx ~]$ uname -a
Linux mrx.internal.excfb.com 2.6.15-1.2054_FC5 #1 Tue Mar 14 15:48:33 EST 2006 i686 i686 i386 GNU/Linux

So, after I unpacked the tarball, Firefox would not start. It said it was already running and if I couldn't kill it, I would have to reboot. Nonsense, I'd never run it on this box. But I had on adept. I just needed to blow away the lock file:

[tdh@mrx ~]$ cd .mozilla/
[tdh@mrx .mozilla]$ cd firefox/
[tdh@mrx firefox]$ ls -la
total 24
drwxr-xr-x 3 tdh wheel 4096 Nov 20 16:35 .
drwxr-xr-x 4 tdh wheel 4096 Jan 15 17:51 ..
drwx------ 7 tdh wheel 4096 Mar 23 10:17 i6cyljlm.default
-rw------- 1 tdh wheel 5071 Mar 22 13:31 pluginreg.dat
-rw-r--r-- 1 tdh wheel  104 Jan 10 23:09 profiles.ini
[tdh@mrx firefox]$ cd i6cyljlm.default/
[tdh@mrx i6cyljlm.default]$ ls -la | grep lock
lrwxrwxrwx 1 tdh wheel      19 Mar 23 00:34 lock -> 192.168.2.108:24627
-rw-r--r-- 1 tdh wheel       0 Mar 23 10:17 .parentlock
[tdh@mrx i6cyljlm.default]$ rm lock