[mirror-admin] Still having problems syncing

[Carlos Carvalho]

Chris Adams (cmadams at hiwaay.net) wrote on 15 July 2009 21:59:
 >If the timestamps have been updated, all the mirrors would be trying to
 >redownload a large amount of data simultaneously, which could be a cause
 >of load issues.

That's what I said yesterday, with a table showing that most of the
churn is just in the timestamp. It all started around the weekend or
at most 2-3 days earlier. I noticed it by a big difference in the
ratio of literal/matched data shown by rysnc. I just checked that
there are no repeated files in the last days runs. Hopefully "some
day" the file supply will end and the churn will stop...

--