Make sure you're signed up as a mirror in the shared Fedora/CentOS Mirrormanager instance: https://admin.fedoraproject.org/mirrormanager/ Link to your node/server in mirrormanager: https://admin.fedoraproject.org/mirrormanager/host/2747
Bandwidth: 10000-100000 Mbps (10-100 Gbit, detailed info with the location) Country Location: US (CA 100Gb, WA 100Gb, IL100Gb, TX 100Gb, FL 100Gb, CO10Gb, NJ 100Gb, NY 100Gb), CA (Ontario 10Gb, Quebec 10Gb), AU 10Gb, SG 10Gb, IN 10Gb, Ch 10Gb, KR 10 Gb, JP 10Gb, GB 100Gb, NL 100Gb, DE 100Gb, SE 10Gb Sponor organization (optional): Lónyai Sponsor URL (optional): https://www.lonyai.com/
MirrorManager cannot handle setups like this. Can you provide a location/country specific URL for each location? We can only assign one URL to one country.
Can you also provide rsync access for our crawler. Crawling via HTTP/HTTPS is not very efficient when compared to RSYNC.
Metadata Update from @arrfab: - Issue assigned to carlwgeorge - Issue tagged with: centos-stream, doc, mirrormanager
Okay, I will create some clone of this on this weekend or next week. Do I create one for US and one for Canada? Or do I create for every states? The config and availability will be same for every mirror site. And I'm sorry but I can't provide rsync access. I can't open and forward the rsync port on this servers.
Do I create one for US and one for Canada? Or do I create for every states? The config and availability will be same for every mirror site.
Every mirror needs to specify a country. So country is enough. In addition we try to redirect clients to you based on ASNs and close networks you can additionally specify. All clients in the same ASN will be redirected to you.
What does this mean for our crawler. Our crawler is based in the US. If we are crawling an URL in Europe will the traffic actually go to Europe or be intercepted by the CDN along the way. In case this happens it would make all the crawling useless?
Metadata Update from @carlwgeorge: - Issue priority set to: Waiting on Reporter (was: Needs Review)
bug triaging in infra and releng stand-up : what's the status on this ? can we close it or else ? Thanks :)
Hi,
I started to deploy a new S3 image but I have problem with the mirror creator VM. I would like to get away cheap... but I have only EU servers and contacts. My plan is a us-east based S3 image and I can mirror this to my edge servers. Can you help me rsyncing the Centos and Appstream mirrors to my S3 at US? Another, I need more time.
I am done with some mirror. I would like to start more ~10 like in the init post.
The original one: HTTP: http://humirror.lonyai.com/appstream/ HTTPS: https://humirror.lonyai.com/appstream/ RSYNC: 213.157.100.71::AppStream Bandwidth: 10000-100000 Mbps (10-100 Gbit, detailed info with the location) Country Location: HU
HTTP: http://uscamirror.lonyai.com/appstream/ HTTPS: https://uscamirror.lonyai.com/appstream/ RSYNC: --- Bandwidth: 100000 Mbps (100Gb) Country Location: US (CA)
HTTP: http://semirror.lonyai.com/appstream/ HTTPS: https://semirror.lonyai.com/appstream/ RSYNC: --- Bandwidth: 10000 Mbps (10Gb) Country Location: SE
Sponor organization (optional): Lónyai Sponsor URL (optional): https://www.lonyai.com/
IPv4 Addresses: 213.157.100.71, 213.157.100.66 IPv6 Addresses: ---
It's confusing to have the top level directory named appstream when it contains all of the BaseOS, AppStream, CRB, HighAvailability, NFV, RT, and ResilientStorage repositories. Could this directory be renamed to centos-stream?
appstream
BaseOS
AppStream
CRB
HighAvailability
NFV
RT
ResilientStorage
centos-stream
Note to future self, here are the site links:
I think that this idea is right. I renamed :-)
I've added the CentOS category to each of these hosts with the appropriate URLs. Let us know if in the future you're able to open up rysnc access to semirror or uscamirror so we can scan them for up-to-date-ness more efficiently.
Metadata Update from @carlwgeorge: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Metadata Update from @carlwgeorge: - Issue priority set to: None (was: Waiting on Reporter)
Log in to comment on this ticket.