[Pdns-users] PowerDNS cannot handle the load?

[Stephane Bortzmeyer]

On Wednesday 9 April 2003, at 9 h 9, 
Mike Benoit <mikeb at netnation.com> wrote:


> What was your cache timeout set to? How often did you query the same
> records? 

The defaults:

cache-ttl=20
query-cache-ttl=20
negquery-cache-ttl=60

There was almost no repetition of queries (1000 domains choosen at random from 
a pool of the 150,000 domains in '.fr).

> What are the hardware specs?

A PC with two 2Ghz processors, 1 Gbytes of RAM and a GigaEthernet card. It 
runs the Debian operating system, "unstable" branch, with the Linux kernel 
2.4.20.

> We can easily handle 500+ queries/sec on a
> PIII-866 with a PostgreSQL backend. 

May be the size of the zone is the issue. What is yours? We used a zone of 
150,000 domains.

> If the queries aren't hitting the
> cache, the bottleneck is almost always the database, have you tuned
> Postgres? 

No, but I have almost the same problem with the BIND backend.

> Try querying the exact same record over and over again with PDNS, how
> many queries/sec do you get then? 

This is much better. No more hanging of the server, a performance which is almost the same as BIND9 (identical for the two backends). It means that the cache works fine. But it is not very realistic :-)

DNS Query Performance Testing Tool
Version: $Id: queryperf.c,v 1.1.1.2.2.1 2001/09/27 00:44:26 marka Exp $

[Status] Processing input data
[Status] Sending queries
[Status] Testing complete

Statistics:

  Parse input file:     once
  Ended due to:         reaching end of file

  Queries sent:         4352 queries
  Queries completed:    4352 queries
  Queries lost:         0 queries

  Percentage completed: 100.00%
  Percentage lost:        0.00%

  Started at:           Wed Apr  9 21:37:32 2003
  Finished at:          Wed Apr  9 21:37:33 2003
  Ran for:              0.618410 seconds

  Queries per second:   7037.402371 qps