Hi folks
I’ve been seeing a lot of Rsync errors lately:
@ERROR: max connections (40) reached -- try again later rsync error: error starting client-server protocol (code 5) at main.c(1503) [receiver=3.0.6]
Is there plans to increase the max connections or is the solution rather to get more mirror sites offering rsync? We offer rsync here if other mirrors want to use it for example – might take some load off the primary?
Cheers,
Paul
Paul Stewart Senior Network Architect Nexicom 5 King St. E., Millbrook, ON, LOA 1GO Web: http://www.nexicom.nethttp://www.nexicom.net/ Nexicom - Connected. Naturally.
On 05/21/2014 12:53 PM, Paul Stewart wrote:
Hi folks
I’ve been seeing a lot of Rsync errors lately:
@ERROR: max connections (40) reached -- try again later rsync error: error starting client-server protocol (code 5) at main.c(1503) [receiver=3.0.6]
Is there plans to increase the max connections or is the solution rather to get more mirror sites offering rsync? We offer rsync here if other mirrors want to use it for example – might take some load off the primary?
do you know what machine you hit when you get this error ?
Thanks…
I don’t know for sure but if I do a lookup at this moment from that machine I see:
[root@centos paul]# ping msync.centos.org PING msync.centos.org (67.212.81.83) 56(84) bytes of data. 64 bytes from node1.hostingbreeze.com (67.212.81.83): icmp_seq=1 ttl=57 time=15.8 ms 64 bytes from node1.hostingbreeze.com (67.212.81.83): icmp_seq=2 ttl=57 time=15.8 ms 64 bytes from node1.hostingbreeze.com (67.212.81.83): icmp_seq=3 ttl=57 time=15.9 ms
Does that help?
Thanks,
Paul
On 2014-05-21, 1:21 PM, "Karanbir Singh" mail-lists@karan.org wrote:
On 05/21/2014 12:53 PM, Paul Stewart wrote:
Hi folks
I’ve been seeing a lot of Rsync errors lately:
@ERROR: max connections (40) reached -- try again later rsync error: error starting client-server protocol (code 5) at main.c(1503) [receiver=3.0.6]
Is there plans to increase the max connections or is the solution rather to get more mirror sites offering rsync? We offer rsync here if other mirrors want to use it for example – might take some load off the primary?
do you know what machine you hit when you get this error ?
-- Karanbir Singh +44-207-0999389 | http://www.karan.org/ | twitter.com/kbsingh GnuPG Key : http://www.karan.org/publickey.asc _______________________________________________ CentOS-mirror mailing list CentOS-mirror@centos.org http://lists.centos.org/mailman/listinfo/centos-mirror
I don’t know for sure but if I do a lookup at this moment from that machine I see: 64 bytes from node1.hostingbreeze.com (67.212.81.83): icmp_seq=1 ttl=57
"hostingbreeze.com" is not registered anymore, this is an old RDNS-entry; while the new customer does not care/want to update it.
On 21/05/14 20:58, Paul Stewart wrote:
Thanks…
I don’t know for sure but if I do a lookup at this moment from that machine I see:
[root@centos paul]# ping msync.centos.org PING msync.centos.org (67.212.81.83) 56(84) bytes of data. 64 bytes from node1.hostingbreeze.com (67.212.81.83): icmp_seq=1 ttl=57 time=15.8 ms
Well, if you use msync.centos.org, us-msync.centos.org , or eu-msync.centos.org (as described on the wiki page http://wiki.centos.org/HowTos/CreatePublicMirrors ), you have to realize that the dns setup for those A records will do some round-robin operations (and GeoIP lookup too) so you can hit a mirror and next time it can be another one. That round-robin process doesn't check the load/number of connections on the msync nodes, so we'd like to know which one you were hitting when you had that issue. The centos node hostname is shown in every rsync connection (header/welcome message)
I have to implement some extra checks, as I guess the msync nodes will be overloaded when CentOS 7 will be pushed to mirrors :-)
Cheers,
Thanks for clarifying. So is it reasonable to assume that a well connected North American mirror that has rsync enabled is also serving content for the msync A record? I though for some reason that the msync record resolved to some dedicated boxes just for serving the mirror sites so just curious.
When we do get the error, it does go away for sometimes a day or two and then comes back sometimes. So in essence other than getting the error every so often, our mirror does keep updated :)
Cheers,
Paul
On 2014-05-22, 6:58 AM, "Fabian Arrotin" fabian.arrotin@arrfab.net wrote:
On 21/05/14 20:58, Paul Stewart wrote:
Thanks…
I don’t know for sure but if I do a lookup at this moment from that machine I see:
[root@centos paul]# ping msync.centos.org PING msync.centos.org (67.212.81.83) 56(84) bytes of data. 64 bytes from node1.hostingbreeze.com (67.212.81.83): icmp_seq=1 ttl=57 time=15.8 ms
Well, if you use msync.centos.org, us-msync.centos.org , or eu-msync.centos.org (as described on the wiki page http://wiki.centos.org/HowTos/CreatePublicMirrors ), you have to realize that the dns setup for those A records will do some round-robin operations (and GeoIP lookup too) so you can hit a mirror and next time it can be another one. That round-robin process doesn't check the load/number of connections on the msync nodes, so we'd like to know which one you were hitting when you had that issue. The centos node hostname is shown in every rsync connection (header/welcome message)
I have to implement some extra checks, as I guess the msync nodes will be overloaded when CentOS 7 will be pushed to mirrors :-)
Cheers,
-- Fabian Arrotin gpg key: 56BEC54E | twitter: @arrfab _______________________________________________ CentOS-mirror mailing list CentOS-mirror@centos.org http://lists.centos.org/mailman/listinfo/centos-mirror
-
Christian Freund -
Fabian Arrotin -
Karanbir Singh -
Paul Stewart