Hi,
"eu-msync.centos.org has address 217.115.169.1 2011-07-09 11:29:46 msync.CentOS.org rsync service (centosh2)"
Above server just deleted 6.0 of our mirror, and some other updates as well. Would it make sense to disable rsync from machines that are not yet up2date during a major version update? That way traffic could be greatly reduced. For now i removed --delete from the script, but in my opinion temporary disabling rsync from out-of-date machines would be better, since it can be better controlled.
On Sat, Jul 09, 2011 at 11:35:57AM +0200, Sander Hoentjen - Oxilion wrote:
Hi,
"eu-msync.centos.org has address 217.115.169.1 2011-07-09 11:29:46 msync.CentOS.org rsync service (centosh2)"
According to Ralph the underlying configuration issue with h2 has been corrected and it's syncing now.
John
Am 09.07.11 11:35, schrieb Sander Hoentjen - Oxilion:
Hi,
"eu-msync.centos.org has address 217.115.169.1 2011-07-09 11:29:46 msync.CentOS.org rsync service (centosh2)"
Above server just deleted 6.0 of our mirror, and some other updates as well. Would it make sense to disable rsync from machines that are not yet up2date during a major version update? That way traffic could be greatly reduced. For now i removed --delete from the script, but in my opinion temporary disabling rsync from out-of-date machines would be better, since it can be better controlled.
This one (I accidentally set it to sync from itself) and the y3 issues seem to be the only issues encountered so far. The other mirrors have their set of data.
h2 now is complete, y3 is still copying. Both problems only have arisen when syncing the dvd tree, so normal msync.centos.org syncs didn't show this behaviour.
Ralph
----- Original Message ----- | Hi, | | "eu-msync.centos.org has address 217.115.169.1 | 2011-07-09 11:29:46 | msync.CentOS.org rsync service (centosh2)" | | Above server just deleted 6.0 of our mirror, and some other updates as | well. Would it make sense to disable rsync from machines that are not | yet up2date during a major version update? That way traffic could be | greatly reduced. For now i removed --delete from the script, but in my | opinion temporary disabling rsync from out-of-date machines would be | better, since it can be better controlled.
This is precisely why I sent a notice that you remove any --delete flags. Any time there is a major release 3=>4=>5=>6 you should remove these flags so that content isn't destroyed while there is a lot of mirror flux. Usually one or two weeks after release I tend to put it back.
-
James A. Peltier -
John R. Dennison -
Ralph Angenendt -
Sander Hoentjen - Oxilion