[tz] Blocked USNO sites was: Updated Public Domain leapseconds.list
Brian.Inglis at SystematicSw.ab.ca
Sat Jul 6 20:04:59 UTC 2019
On 2019-07-06 11:07, Steve Allen wrote:
> On Sat 2019-07-06T10:12:26-0600 Brian Inglis hath writ:
>> Seems to be some DNS resolver and network blocking, possibly only outside the
>> US, to maia.usno.navy.mil and www.usno.navy.mil; interactive web browser access
>> to tycho.usno.navy.mil and aa.usno.navy.mil mostly work
> "mostly work" is consistent with our recent robotic telescope experience.
> We have a longstanding procedure to retrieve Bulletin A every week.
> In the past couple of months that has been failing, so we have
> modified the script to try more than once.
Maybe we need not be too paranoid then about being alien data users.
They give an example script and README:
using maia.usno.navy.mil as primary and toshi.nofs.navy.mil as secondary.
I would follow that example but switch primary and secondary, as toshi has been
more accessible and reliable in recent months (so far, fingers crossed), reduce
time spent trying each server, and try to get newer data from any usable servers.
When there have been access and reliability problems in the past, as with NIST
FTP and other mirror lists, it usually works better and faster to reduce
wget -T timeout (900s) and -t tries (20) to low single digits, and cycle thru
the mirrors list, with short sleeps between servers, and longer sleeps between
cycles, dropping candidates with DNS or connection problems (as they are
unlikely to clear up quickly).
If the candidates list gets emptied, it gets reset from the mirrors list.
The cycle continues until all required data files have been retrieved for
checking or updating, or cycle count or time limits are reached.
I tend to limit trying and failing, and let tomorrow's or next week's run deal
Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada
This email may be disturbing to some readers as it contains
too much technical detail. Reader discretion is advised.
More information about the tz