Does the White Russian 0.9 DynDNS client suck just as much?

[Ben Scott]

On 5/14/07, VirginSnow at vfemail.net <VirginSnow at vfemail.net> wrote:
> OpenWRT was recommended as a way of getting around using Linksys's
> broken DynDNS client.  But this system seems just as broken!

  I suspect something *is* broken in the OpenWRT DDNS subsystem.  But
I was able to get the symptom (i.e., DynDNS warning about my expiring
DNS entry) to go away by fiddling with it.  I've been too lazy/busy to
dig into root cause yet.  I wasn't sure if the problem wasn't just me
until now.  I guess it's not just me.

  Anyway, here's what I did.  I logged into a root shell on the
LinkSys box.  I edited/created the /etc/ez-ipupdate/ez-ipupdate.conf
file to look like this:

interface=vlan1
service-type=dyndns
user=notmyusername:notmypassword
pid-file=/var/run/ez-ipupdate.pid
cache-file=/etc/ez-ipupdate/ez-ipupdate.cache
host=notmyhostname.dnsalias.org

  You can probably guess that I've changed the "user" and "host"
directives to be generic; change to match your situation.

  Once I did this, I did "killall ez-ipupdate" (hi, Solaris admins!),
then "/etc/init.d/S52ez-ipupdate start", and since then, it's always
worked.  I know I've had power failures since then, and I also know
the unit isn't on a UPS (since the UPS for that room died a while back
and I haven't replaced it yet), so the "fix" apparently lives through
a power loss.

  As an aside, one major reason I'm so enthused about OpenWRT is that
one *can* fiddle with it.  It's not that LinkSys's software is
inherently broken and that OpenWRT is inherently not-broken.  It's a
lot of the same code.  The difference is that LinkSys welds the hood
shut, while OpenWRT at least gives me the chance to do something about
it.

  (The other major reason I'm enthused about OpenWRT is the spiffy
real-time traffic graphs in X-WRT.  Sooooo cool.  :)  )

> And there isn't any button labeled "forced update".

  You can also run "ez-ipupdate" from the shell prompt.  Give it a
"--help" argument for some fairly useful help.

> # date; uptime
> Mon May 14 09:18:51 EDT 2007
>  09:18:51 up 21:16, load average: 0.00, 0.00, 0.00
>
> I didn't reboot the box at all yesterday, so the 21:16 is wrong nonetheless.

  I suspect the uptime is correct, and the box has only been up for 21
hours.  The real question is: Why did the box reboot?  The nominal
possible causes are: User action, power problems, software fault,
hardware fault.  We can rule out user action, since you didn't do
anything (right?).  Can you comment on power reliability at your
installation?  If power isn't the cause, either the software is
crapping out and leading to a reboot (in which case that's a symptom
of something needing fixing), or the hardware is failing (in which
case all bets are off anyway).  So this may be worth investigating.

  One thing you can do is, if you've got another *nix machine on your
LAN up all the time, is setup a syslog listener on that other machine,
and configure OpenWRT to forward all the syslog events to that
machine.  Then you can have those messages logged permanently, even if
the LinkSys box crashes or resets.

  On a mostly unrelated note, DynDNS emails me when my hostname is
about to expire, so I've never lost the name due to any issues with
LinkSys's firmware, OpenWRT's firmware, or anything else.  I've always
had a chance to tweak things first.  You may want to look into getting
the same email warning for yourself.

-- Ben