12-28-2010 01:23 PM - edited 12-28-2010 01:24 PM
My 650 cluster is configured and so far appears to be running properly. I can ssh to the management port of both devices and ping from each back to the ntp server.
When I run ntpdate -u timeserver from the secondary device I get a response and the time is synchronized.
tcpdump on the time server shows ntp requests coming from the primary but (except when running ntpdate) I never see ntp requests from the secondary.
Is this normal or is there something I can add to my config to get the secondary to periodically request sync?
01-04-2011 07:26 AM
This KB article kind of hints that it can't be done for some reason:
"You can specify that the system time is retrieved from the NTP server when the device boots or enters a chassis cluster backup state"
Otherwise take a look in the log for the backup device and see if anything stands out: show log messages | match ntp
01-04-2011 10:51 AM
set system ntp boot-server [address-of-ntp]
set system backup-router [address-of-nexthop-on-fxp0] destination [network-that-ntp-and-mgmt-is-in]
That should, hopefully, do it. The backup device doesn't have an active RE, which means it needs to get to the NTP server via fxp0. That requires a boot-server to be set as well as a backup-router.
Do let us know whether that works out. If not, and the boot-server is indeed only used at boot, then I suppose a jscript would be the only alternative.