Hi all,
I need your opinion regarding performance and stability of DJBDNS as compared to BIND 9. I am planning to give djbdns a chance to prove itself on a pretty good hardware configuration (3.2 GHz Xeon, 2 GB DDR RAM, 2 x 72.3 SCSI). Its an HP ML 450 to be specific and will run CentOS 4.2. your feedback will be highly appreciated.
Regards
~zaeem
Bind 9 vs DJBDNS
-
- Lieutenant Colonel
- Posts: 660
- Joined: Sat Jul 06, 2002 12:35 pm
- Location: Islamabad
- Contact:
Re: Bind 9 vs DJBDNS
djbdns is stable -- i've never had any trouble with it.
if you're going to use it for caching, give the cache lots of memory, and use this to tune it. if you're going to use it to serve domain information, make sure you write up lots of notes so the next sysadmin who comes along to replace you knows how to deal with it.
if you're going to use it for caching, give the cache lots of memory, and use this to tune it. if you're going to use it to serve domain information, make sure you write up lots of notes so the next sysadmin who comes along to replace you knows how to deal with it.
-
- Lieutenant Colonel
- Posts: 660
- Joined: Sat Jul 06, 2002 12:35 pm
- Location: Islamabad
- Contact:
oh yeah? well, try this!
-
- Lieutenant Colonel
- Posts: 660
- Joined: Sat Jul 06, 2002 12:35 pm
- Location: Islamabad
- Contact:
That is great. Any tool which can reliably stress test both Bind and DJB DNS? I have a HP ML 450 with 2 GB RAM and I want the maximum use of it. A stress tool that can test both authoritative and cache query performance on both these servers?
As a side note, I am not completely satisfied with the way BIND 9 uses available resources and that is the main reason for trying other solutions.
~Zaeem
As a side note, I am not completely satisfied with the way BIND 9 uses available resources and that is the main reason for trying other solutions.
~Zaeem
i searched, but didn't find anything useful. djbdns ships with dnsfilter, which tests dnscache primarily. i suppose if the ip addresses you check with dnsfilter are in a zone managed by your tinydns, you can test both services with it.
one thing i can recommend is cutting down on the lines logged by dnscache. read the docs for multilog, and change /service/dnscache/log/run so that it doesn't log lines beginning with "rr", "cached", or "tx".
but really, with the machine you have, you'll probably saturate your uplink long before djbdns starts slowing down.
one thing i can recommend is cutting down on the lines logged by dnscache. read the docs for multilog, and change /service/dnscache/log/run so that it doesn't log lines beginning with "rr", "cached", or "tx".
but really, with the machine you have, you'll probably saturate your uplink long before djbdns starts slowing down.