I think I need an second, third, etc set of eyes on this issue.
In design, we have had our external DNS servers located in our DMZ poriton of PIX (three interfaces). With this implimentation, we have had no issues (WIN2K DNS).
With some business need changes and so forth, we moved the DNS from WIN2K to Red Hat Linux and BIND 9.x
After this move, we did no experience any issues for about a week and a half. Then it happened....
NO MORE DNS RESOLUTION!
According to PIX itself, it logged UDP traffic to port 53 with no issues. Checked the syslog, all looks good. Then we looked at the DNS traffic to DNS server...it was not getting any traffic at all, even though PIX said that it pushing it through.
In testing (and in tryong to get back up as the assumption was the DNS server and maybe a NIC issue) we put the WIN2K DNS server back into place...still no go..same symptoms.
With this, we placed the WIN2K DNS outside PIX and all was well, then we moved the Linux box out, and all was well. WITH NO CHANGES TO SERVER HARDWARE.
With that I defer to the expert crowd...I have opened a TAC case, but they stabbing in the dark on this one.
I have a static mapping for my DNS box normally, and an ACL that looks like this...but it has worked forever!
access-list outside_in permit udp any host {ip} eq domain
Cisco is telling me that there maybe an issue due to the fact that I have built an outbound filter for internal clients (to get rid of all the IM crap on my network). I do not buy into this as the filter been apart of the orginal config.
Any ideas?
ilurec
In design, we have had our external DNS servers located in our DMZ poriton of PIX (three interfaces). With this implimentation, we have had no issues (WIN2K DNS).
With some business need changes and so forth, we moved the DNS from WIN2K to Red Hat Linux and BIND 9.x
After this move, we did no experience any issues for about a week and a half. Then it happened....
NO MORE DNS RESOLUTION!
According to PIX itself, it logged UDP traffic to port 53 with no issues. Checked the syslog, all looks good. Then we looked at the DNS traffic to DNS server...it was not getting any traffic at all, even though PIX said that it pushing it through.
In testing (and in tryong to get back up as the assumption was the DNS server and maybe a NIC issue) we put the WIN2K DNS server back into place...still no go..same symptoms.
With this, we placed the WIN2K DNS outside PIX and all was well, then we moved the Linux box out, and all was well. WITH NO CHANGES TO SERVER HARDWARE.
With that I defer to the expert crowd...I have opened a TAC case, but they stabbing in the dark on this one.
I have a static mapping for my DNS box normally, and an ACL that looks like this...but it has worked forever!
access-list outside_in permit udp any host {ip} eq domain
Cisco is telling me that there maybe an issue due to the fact that I have built an outbound filter for internal clients (to get rid of all the IM crap on my network). I do not buy into this as the filter been apart of the orginal config.
Any ideas?
ilurec