Cacti problem when 1st DNS server in resolv.conf is offline

[jdawg]

Hello,

I'm having a frustrating time trying to figure this out. I am running Cacti 0.8.7g on Centos 5.5 and have 2 DNS servers listed in resolv.conf. Cacti has worked great for over a year until today. This morning, the first(primary) DNS server listed in resolv.conf failed and went offline. It appears that Cacti did not use the 'failover' or secondary DNS server listed as it should have, but was stuck on the first one. From the CentOS OS, I was able to connect to resourses via fully qualified domain name, so I know that applications like ssh and telnet were able to use the resolv.conf correctly, since the 2nd listed DNS was used. Also, sendmail from the command line worked.

However, Cacti was unable to access resources via FQDN when the first DNS server listed in resolv.conf was unavailable. I was able to reboot the failed DNS server several hours later and Cacti started collecting data and graphing again. Is there a setting somewhere that forces Cacti to use a DNS server(or cache DNS settings/responses)? Thanks for everyone's help.

John

[noname]

Try to increase "SNMP Timeout" (per device settings) and/or "SNMP Retries" (in 'Console -> Settings').

[gandalf]

Cacti does not use anything specific regarding DNS. It simply calls gethostbyname or the like.
But making use of DNS timeouts to failover to the second DNS is not a good move. The DNS timeouts are way too long for mass polling. And in case you increase host timeouts to be >= DNS timeouts, I fear you won't be happy at all.
I suppose it would be better to solve the issue DNS-wise
R.

[Christo]

Hi,
Dear all i think i can help you some thing please visit this link.

http://www.techyv.com/article/how-impro ... dns-server