On Feb 3, 2:21 pm, Harlan Stenn <[email protected]> wrote: > >>> In article > >>> <2eea3363-767e-44f4-8d85-17d9838d6...@r28g2000vbp.googlegroups.com>, > >>> SteveInTallyFL <[email protected]> writes: > > SteveInTallyFL> Can one of you experienced NTP folks please verify the > SteveInTallyFL> reasonableness and safety of this time plan. Thanks. > > SteveInTallyFL> 2 servers in my DMZ and 1 internal server getting time from > SteveInTallyFL> 0 - 4.us.pool.ntp.org and peered with each other > > SteveInTallyFL> 4 internal servers in an Oracle Real Application Cluster, > SteveInTallyFL> each pointing to these 3 three servers, also peered to each > SteveInTallyFL> other. > > SteveInTallyFL> Question: If these four cluster servers are getting time > SteveInTallyFL> from the three peers is there anything gained or lost by > SteveInTallyFL> peering them to each other? > > I like having client server boxes peer with each other because it makes it > real easy to see how things are doing on just by running 'ntpq -p'. > > This may be less important if you have comprehensive monitoring of all your > machines. > > -- > Harlan Stenn <[email protected]>http://ntpforum.isc.org - be a member!
Monitoring of the cluster and the application server (which is also a time server source) is limited. We currently are having application- level issues because the cluster members are getting outside an acceptable jitter range among each other and with the application server. My take-away from your response if that this is a good thing to do. Thank you for feedback. _______________________________________________ questions mailing list [email protected] https://lists.ntp.org/mailman/listinfo/questions
