From: Matt Lawrence <matt@...>
Date: 2007-01-23T08:56:20+09:00
Subject: Re: resolv-replace problem

On Tue, 23 Jan 2007, Eric Hodel wrote:

> On Jan 22, 2007, at 14:44, Matt Lawrence wrote:
>> After running through aproximately 512 ip to name resolutions, the 
>> resolv-replace library seems to become very slow and starts blocking 
>> threads.  Any ideas?
>> 
>> Any good alternatives for doing ip address to name lookups on thousands of 
>> systems?
>
> I've experienced this problem and assumed it was due to resolv-replace 
> waiting for the destination name servers to time out, but didn't verify. 
> Have you examined this possibility?
>
> If you've got 1000 bogus host names, 100 threads and a 30 second timeout per 
> thread you'll still wait at least 5 minutes for all of them to fail to 
> resolve.

Shouldn't have to wait for all of the lookups to fail.  Every time EACH 
lookup fails, the threads should continue.  I have a limit in my program 
of 50 threads.  After about 512 lookups, the number of threads goes from 
50 active to 6 or 7 active.  Even when using the same address range over 
and over.

-- Matt
It's not what I know that counts. 
It's what I can remember in time to use.