Not reducing the bandwidth means that the affected servers (like mine) drop out of the pool all together. By reducing the bandwidth I was able to get my server back in the pool at least.
Regards Austin On 23 May 2015 at 10:24, Denis Corbin <[email protected]> wrote: > On 22/05/2015 20:47, Maximiliano Valdez wrote: > > Matt, sorry for the duplicity, I forgot to reply to the list > > Hi all, > > > Hi Matt > > > > I wonder if this behaviour is the result of some kind of load balancing > > error on ntp.org DNS or something. > > > > An DoS attack using IP source spoofing could cause the same behavior > from the attacked server point of view. Difficult to know whether this > is a load balancing error or an single attack. > > However, if the NT pool is reduced for Brazil due to some network issues > for example, the behavior is different: all server still remaining in > the pool should see the same load increase, well, more or less. > > The number of active server in the Brazil pool is available at > http://www.pool.ntp.org/zone/br > > First there is a small number of servers and load balancing may > statistically be unfair more often than for a larger pool. Second, the > number of active server dropped from 20 down to 9 since the beginning of > the year. This alone could explain the load increase Matt observes on > his server. > > Reducing the bandwidth in "Manage your servers" will increase the load > the other servers which might lead their respective maintainers to > reduce the bandwidth allocation at the same place which will increase > back the load on your server... > > Cheers, > Denis. > > _______________________________________________ > pool mailing list > [email protected] > http://lists.ntp.org/listinfo/pool > _______________________________________________ pool mailing list [email protected] http://lists.ntp.org/listinfo/pool
