NSLU2 keeping time (badly)

[Ted Harper]

While watching the Firefly logfiles on my NSLU2, I noticed how incredibly badly it seems to keep time. It is running the latest unSLUng firmware (6.8 beta) and has not been de-underclocked, and seems to have managed to lose over 30 minutes in the 3 days since I set it up.

Anyway, I googled a bit and have installed the ntpclient package, and changed the hourly clock-setting event in /etc/crontab from the default to be instead:

1 * * * * root /opt/bin/ntpclient -s -l -h 0.pool.ntp.org > /var/log/ntp.log && hwclock –systohc –utc

This seems to work (as far as I can tell so far). Has anyone done anything different from this? The above seems to work ok and avoids having another daemon running. While testing the various Firefly nightlies if I don’t have the clock on the NSLU2 at least close to correct, it is really hard to correlate funny messages in the log to things I might be doing on the Soundbridge.

ted.h.

[rpedde]

@Ted Harper wrote:

1 * * * * root /opt/bin/ntpclient -s -l -h 0.pool.ntp.org > /var/log/ntp.log && hwclock –systohc –utc

What I’m doing is substantially the same. I’m using ntpdate or something, but still substantially the same.

And yeah. Talk about bad clock…

From an “obtw” standpoint, deunderclocking takes no time and it actually provides some pretty reasonable benefit. I run two slugs — on de-underclocked running debonaras (debian everywhere!) and the other running unslung underclocked (as a slow testbed).

I can definitely tell the deunderclocked one. And not just because it’s running a sweet distro. 🙂

[Author]

Posts