It is very solid.  We are using *all* of their regions.  As a result of this, 
we've been able to pin point that there are only a couple countries, *in the 
world* that *don't* use ClamAV.  It's very impressive.


On Jul 10, 2018, at 10:13 PM, Eric Tykwinski 
<[email protected]<mailto:[email protected]>> wrote:

They have some documentation on their site: 
https://support.cloudflare.com/hc/en-us/articles/115000540888-Load-Balancing-Geographic-Regions
No clue what regions they are using, but hopefully they donated some, it’s a 
pretty solid anycast system.

Sincerely,

Eric Tykwinski
TrueNet, Inc.
P: 610-429-8300

On Jul 10, 2018, at 10:03 PM, Freddie Cash 
<[email protected]<mailto:[email protected]>> wrote:

Joel posted pictures (in one of these update thread) of where the mirrors are 
located along with the relative traffic that each one transfers.

Cheers,
Freddie

Typos courtesy of my phone's keyboard.

On Tue, Jul 10, 2018, 6:37 PM Paul Kosinski, 
<[email protected]<mailto:[email protected]>> wrote:
I have a question. I presume that there are more physical Cloudflare
server instances than implied by 
database.clamav.net<http://database.clamav.net/>'s 5 IP addresses,
and that they are geographically distributed, rather than all being
in/near San Francisco. This suggests that they are Anycast addresses.
But I don't know how to determine where the server instances are
located, or which one(s) we reach when trying to download cvds.

The fact that we have observed a 1 hour delay further suggests that
there a large number of instances, otherwise they would be brought into
sync with the DNS TXT record more quickly. Is there any way that you
people at ClamAV can determine when the various server instances in fact
get the new cvd files? I would think that a CDN would provide statistics
on that, especially if expected delays are spelled out in an SLA.


On Tue, 10 Jul 2018 22:11:46 +0000
"Joel Esler (jesler)" <[email protected]<mailto:[email protected]>> wrote:

> Thanks for this feedback everyone.  This is extremely useful.
>
>
> > On Jul 10, 2018, at 11:26 AM, Paul Kosinski
> > <[email protected]<mailto:[email protected]>> wrote:
> >
> > Last night our new method of getting cvd updates showed that it was
> > *one hour* from the time the DNS TXT record claimed a new cvd was
> > available to the time when our quick curl said it was really
> > available!
> >
> > In particular at 1:03 AM (EDT), DNS said version 24739 was
> > available, but a curl of the first few bytes of the cvd file said
> > it was still at version 24738. It wasn't until 2:03 AM that curl
> > reported that version 24739 was really available for download.

>
_______________________________________________
clamav-users mailing list
[email protected]<mailto:[email protected]>
http://lists.clamav.net/cgi-bin/mailman/listinfo/clamav-users


Help us build a comprehensive ClamAV guide:
https://github.com/vrtadmin/clamav-faq

http://www.clamav.net/contact.html#ml
_______________________________________________
clamav-users mailing list
[email protected]<mailto:[email protected]>
http://lists.clamav.net/cgi-bin/mailman/listinfo/clamav-users


Help us build a comprehensive ClamAV guide:
https://github.com/vrtadmin/clamav-faq

http://www.clamav.net/contact.html#ml

_______________________________________________
clamav-users mailing list
[email protected]<mailto:[email protected]>
http://lists.clamav.net/cgi-bin/mailman/listinfo/clamav-users


Help us build a comprehensive ClamAV guide:
https://github.com/vrtadmin/clamav-faq

http://www.clamav.net/contact.html#ml

_______________________________________________
clamav-users mailing list
[email protected]
http://lists.clamav.net/cgi-bin/mailman/listinfo/clamav-users


Help us build a comprehensive ClamAV guide:
https://github.com/vrtadmin/clamav-faq

http://www.clamav.net/contact.html#ml

Reply via email to