Hi Michael,

thanks for the detailed report. I see the problem. The thing is, GU is a small 
country without mirrors, and MirrorBrain doesn’t know that it is better 
connected to US mirrors than to Asian/Australian mirrors. MirrorBrain looks for 
mirrors in the same country first (there aren’t any in Guam), then in the 
continent (Ocenania has a few, but they are slow, as you mention), so it falls 
back to geographic coordinates, which yields the mirrors in Japan, Australia, 
New Zealand, which are not well connected to Guam, too.

A fix is possible. However, I can’t add it in software, but it needs to be 
added by operators of mirror networks, i.e. administrators running MirrorBrain 
on their origin/redirecting server. 

-> Dear Hostmaster at TDF, please add ‚gu‘ to the otherCountries field to some 
US mirrors, especially those which Micheal recommends. (This is done using the 
‚mb edit <mirrorname>‘ command.

I think that for special cases like this it would be good to hardcode the 
„knowledge“ into MirrorBrain, or at least offer some standard settings as a 
default. It’s just a matter of (per-mirror-) configuration, you see. 

PS. One of those countries I never heard of. I entered „Guam“ into my maps app 
this morning, saw an island. Zoomed out, zoomed out, zoomed out, .. and was 
amazed. Wonderful! That’s why MirrorBrain exists!

PPS. In the past, I configured a few German mirrors (which are usually well 
reachable) for similar cases.
otherCountries : dk,pl il,tr,vn 
 a2 ge,kz,ru

PPS. There seems to be good connectivity from most places to the US, and to 
Germany, but it is wrong to assume that remote countries have good reachability 
to their neighbour countries. The South African countries are a good example 
that I studied in some detail. That’s why mirror selection by geographic 
distance can always only be some fallback mechanism with varying mileage. 

> Am 15.02.2016 um 06:58 schrieb Michael D. Setzer II <mi...@kuentos.guam.net>:
> In Guam, going to the LibreOffice site to download the latest version, the 
> automatic process shows mirros in au, nz, and jp as being closest, but the 
> connection speeds really really suck. Reports over 1 hour to download. 
> Manually, pickin a US site, download happens in about 1 minute or 60 times 
> faster. The local ISP has setup the site to look like a US site, since that  
> is 
> where the bigger pipes from Guam go. While physically AU, NZ and JP are 
> closer the traffic seems to have to go to the US anyway to then go the the 
> closer locations. 
> Its a pain to have to manually pick mirrors, and on some sites, that isn't an 
> option. 
> Mirrors
> List of best mirrors for IP address, located at 
> 13.444300,144.786301 in Guam (GU), network 
> (autonomous system 3605).
> Map showing the closest mirrors
> Found 2 mirrors in other countries, but same continent (OC)
> http://mirror.aarnet.edu.au/pub/tdf/libreoffice/stable/5.1.0/rpm/x86_64/LibreOf
> fice_5.1.0_Linux_x86-64_rpm.tar.gz (au, prio 100)
> http://www.ftp.ne.jp/office/tdf/libreoffice/stable/5.1.0/rpm/x86_64/LibreOffice_
> 5.1.0_Linux_x86-64_rpm.tar.gz (au, prio 100)
> Using the default it reports and hour, but manually selecting wayne.edu it 
> takes about 1 minute??
> +----------------------------------------------------------+
>  Michael D. Setzer II -  Computer Science Instructor      
>  Guam Community College  Computer Center                  
>  mailto:mi...@kuentos.guam.net                            
>  mailto:msetze...@gmail.com
>  Guam - Where America's Day Begins                        
>  G4L Disk Imaging Project maintainer 
>  http://sourceforge.net/projects/g4l/
> +----------------------------------------------------------+
> http://setiathome.berkeley.edu (Original)
> Number of Seti Units Returned:  19,471
> Processing time:  32 years, 290 days, 12 hours, 58 minutes
> (Total Hours: 287,489)
> ABC         16613838.513356   |   EINSTEIN    87525353.785695
> ROSETTA     40117323.878181   |   SETI        74413803.416125

mirrorbrain mailing list
Archive: http://mirrorbrain.org/archive/mirrorbrain/

Note: To remove yourself from this mailing list, send a mail with the content
to the address mirrorbrain-requ...@mirrorbrain.org

Reply via email to