Learn more about these different git repos.
Other Git URLs
Hi,
operations on remote repositories ("git remote update", "git pull", "git push") take a long time. It means >2 minutes for operation. Issues are seen for example on fedpkg, rpkg and pungi repositories and this state still takes for about 3 weeks. Protocol I use is ssh. Commands will end correctly, it just takes time. I will check it also from a different place (not just in office) to exclude potential local network issues.
What do you mean with remote repositories? Repositories used for remote pull-requests?
Sorry for not beeing precise. Clone for example ssh://git@pagure.io/rpkg.git And try git remote update or git pull in the cloned repo. Even clone itself takes more than 2 minutes.
git remote update
git pull
We need more information like where you are located, what the route it is taking to get to pagure.io etc.
OK, currently Red Hat office in Brno (Czech Republic). If you need more specific information like traceroute output (it has 25 hops), let me know. Interesting is, that web interface is good in response. Additionally, it seems there is no difference if I am connected through company VPN (BRQ) or not.
traceroute
See if there is a difference by putting in /etc/hosts 140.211.169.204 pagure.io www.pagure.io
Currently it tries to get there throught IPv6:
ping pagure.io PING pagure.io(pagure01.fedoraproject.org (2605:bc80:3010:600:dead:beef:cafe:fed8)) 56 data bytes From 2620:52:0:2200::fe (2620:52:0:2200::fe) icmp_seq=1 Destination unreachable: Address unreachable
So your workaround is working for me:
ping pagure.io PING pagure.io (140.211.169.204) 56(84) bytes of data. 64 bytes from pagure.io (140.211.169.204): icmp_seq=1 ttl=46 time=196 ms
... and response is sufficient now. I do not know the reason of this behaviour. Thanks for your suggestion!
OK can you do a mtr -6 from your host to pagure01.fedoraproject.org ? We have seen several problems where ipv6 in Europe just does not route to some US domains usually many hops beyond anything we can 'influence' by opening tickets.
mtr -6 pagure01.fedoraproject.org
My traceroute [v0.92] onosek (2620:52:0:2200:56e1:adff:fe81:a42f) 2019-02-21T13:50:23+0100 Keys: Help Display mode Restart statistics Order of fields quit Packets Pings Host Loss% Snt Last Avg Best Wrst StDev 1. 2620:52:0:2200::fe 0.0% 36 1.3 1.7 1.3 4.6 0.6 2. ???
It looks normal I think. But git query through IPv6 takes long time. IPv4 is OK.
No that mtr means it got one hop out and never got any further.. so the delays you are seeing is that the system is trying ipv6, timeing out and then going for ipv4. Forcing the name in /etc/hosts short circuits this timeout you are experiencing.
Ok, thanks for the explanation.
Going to close this out as it seems fixed.
Metadata Update from @smooge: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Login to comment on this ticket.