Forum Discussion

Graham_Dobson_1's avatar
Graham_Dobson_1
Icon for Nimbostratus rankNimbostratus
May 30, 2014

LTM Active/passive cluster clocks wrong

Hi,

 

I've inherited a couple of LTMs (BIGIP 1600) in a cluster and whilst everything appears to work - I noticed their clocks are out according to the GUI.

 

both LTMs are configured for an NTP server but it does not appear to be working!

 

The current active node is: 11:42 AM (GMT)

 

The current standby node is: 12:23 PM (GMT)

 

The current time (as I write this) is 12:47 PM (BST)

 

Config Sync appears to be working fine.

 

Obviously - they are both GMT as well, should they be switched to BST? (or GB as its listed)

 

The other question around the whole topic is if I switch the LTMs to using an NTP server that works, and or change the Timezone, am I going to break the clustering?

 

Whats the safest way to get the clocks back in sync and get the timezones set correctly - without breaking everything on this production cluster.

 

Thanks

 

Graham

 

application delivery
deployment
LTM
management

6 Replies

Sort By
  • Cory_50405's avatar
    Cory_50405
    Icon for Noctilucent rankNoctilucent
    May 30, 2014

    Try running 'ntpq -pn' on both of your LTMs and see if they are successfully pulling time from your NTP server.

     

    Setting a new NTP server or changing timezone isn't going to break your active/standby pair.

     

    • Graham_Dobson_1's avatar
      Graham_Dobson_1
      Icon for Nimbostratus rankNimbostratus
      Jun 02, 2014
      Hi, its not so much changing the config that I'm worried about breaking things - its the system times changing by large amounts that might cause problems. What about things like cookie persistance? would that be affected by the system time changing by a large amount? (we're currently loadbalancing an exchange 2010 envirnoment) Graham
  • Domai's avatar
    Domai
    Icon for Altostratus rankAltostratus
    May 30, 2014

    Also you can try to recycle your ntp deamon on the ltms. It helped me before.

     

  • Graham_Dobson_1's avatar
    Graham_Dobson_1
    Icon for Nimbostratus rankNimbostratus
    Jun 02, 2014

    An ntpq -pn gives the following:

     

    [xxxxx@bigip2:Active:In Sync] etc ntpq -pn

     

    remote refid st t when poll reach delay offset jitter

    *x.x.x.7 x.x.x.11 3 u 186d 256 0 0.606 1.271 0.000 [root@bigip2:Active:In Sync] etc date Mon Jun 2 13:57:33 GMT 2014

     

    ...the time on the BIG-IP does not seem to get in sync

     

    Graham

     

  • Cory_50405's avatar
    Cory_50405
    Icon for Noctilucent rankNoctilucent
    Jun 02, 2014

    Your 'when' field is 186d, meaning your BIG-IP hasn't received a packet from your NTP server in 186 days. Can you verify connectivity to your NTP server?

     

  • Graham_Dobson_1's avatar
    Graham_Dobson_1
    Icon for Nimbostratus rankNimbostratus
    Jun 02, 2014

    Hmm...this is odd - I also noticed the "reach" reference on the output was 0 - meaning its unreachable.

     

    for starters I decided to correct the Timezones, from GMT to GB (BST) which I did on the active node - as soon as I did that and clicked update - the times sync'd up perfectly, and the reach is now not 0 and the "when" is 1

     

    ?

     

    Graham