[OSM-dev] Planet .osm diffs

Jon Burgess jburgess777 at googlemail.com
Wed Jan 21 15:32:21 GMT 2009


2009/1/21 Tom Hughes <tom at compton.nu>:
> Mathieu Arnold wrote:
>>
>> +-le 21.01.2009 13:29:10 +0000, Tom Hughes a dit :
>> | Mathieu Arnold wrote:
>> |
>> |> It's been about 10 hours since the planet diffs went out of sync, it seems
>> |> that the box's clock is 30 minutes behind, as of now, the last generated
>> |> diff is 200901211215-200901211216.osc.gz 21-Jan-2009 12:21  7.4K, and it's
>> |> 12:52 UTC...
>> |
>> | Well if I know which box you were talking about (we have quite a lot)
>> | then I could have a look at why it's clock is not in sync.
>>
>> Well, I don't know, I'm talking of what can be found at :
>> http://planet.openstreetmap.org/minute/
>
> I assume it was dev, but I think Jon got to it before me as the clock
> was right when I looked at it. The ntp daemon wasn't running though so I
> have restarted that.

I logged into the machine to free up some disk space at around 1AM
this morning and noticed the clock was 150s slow. It looked like one
of the two UCL NTP servers was maybe returning bogus data because it
thought it was synced even though it was wrong. I stopped ntpd, ran
ntpdate to fix the time, then started ntpd again. Maybe the process
decided to quit when it found the two NTP servers were inconsistent.

I also noticed that the lm_sensors stuff is currently broken on dev.
The munin temperature stats have not been updating since October.

-- 
    Jon




More information about the dev mailing list