06-27-2014 03:27 PM
Dear All,
I am working on a cluster configuration for netbackup services running in two different sites (global/metro cluster).
1) The issue i am facing is that each site has different ip address segment, for example, when the master active node is running on site 1 the master ip address is 1.0.0.1 furthermore when the active node change to second site the ip address is 2.0.0.1
So to identify the active ip addres for the virtual name we are using the VCS Agent for DNS in order to update the right IP on the DNS server.
Because the clients looks for the ip address from the DNS, they can communicate via ping but when we run bpclntcmd -hn master_server it returns the old ip address ( ip address from passive node that was umplumbed ), so I run the bpclntcmd -clear_host_cache; after that netbackup services run well
Based on the enviroment my question is, are there any option to disable the netbackup host cache in order to resolve the correct ip address?
2) When i run the switchover the media server that was active configured appear as offline mediaserver,
Does any body know if the issue is regarding to the host cache?
Best Regards.
Solved! Go to Solution.
06-28-2014 11:03 PM
Hi,
Your issue is caused by PBX. PBX caches hostname entries but you can, and should clear it with
bpclntcmd -clear_host_cache (run it a fews times and then run bpclntcmd -self).
Then you should also do this for the media servers as their cache would still reflect the primary sites name. If you're running unix/linux and have remote execution (ssh) on all systems you could probably put it all in the preonline script.
You should put this into a pre online script so it runs before the NetBackup starts in either site, or do it manually during switch over to the remote site.
06-27-2014 05:43 PM
Does Any body knows how to clear the host cache remotely to the clients.
Best regards
06-28-2014 02:53 AM
You can change the windows cache time in the OS
Changing DNS Cache Settings in Windows
http://www.tech-faq.com/how-to-flush-dns.html
06-28-2014 10:24 AM
Hello.
As read in technote: ( http://www.symantec.com/docs/TECH179465 )
NetBackup 7.0.1 and above cache DNS provided hostnames and IP addresses with a time-to-live (TTL) of 3600 seconds. Depend on when the DNS IP Failover occurs, it may be up to an hour before the entry is refreshed. Once the refresh occurs, normal NetBackup operations will resume.
A delay of up to an hour before resuming successful operations may be acceptable for some hosts at some sites, but not for others. The delay can be minimized by shortening the host cache TTL on critical host. Please review the information below carefully before proceeding
Regads
06-28-2014 11:03 PM
Hi,
Your issue is caused by PBX. PBX caches hostname entries but you can, and should clear it with
bpclntcmd -clear_host_cache (run it a fews times and then run bpclntcmd -self).
Then you should also do this for the media servers as their cache would still reflect the primary sites name. If you're running unix/linux and have remote execution (ssh) on all systems you could probably put it all in the preonline script.
You should put this into a pre online script so it runs before the NetBackup starts in either site, or do it manually during switch over to the remote site.