[bind10-dev] b10-auth stops responding for several tens of seconds after receiving IXFR

JINMEI Tatuya / 神明達哉 jinmei at isc.org
Wed Jan 9 16:17:45 UTC 2013


At Wed, 09 Jan 2013 18:29:30 +0900 (JST),
Naoki Kambe <kambe at jprs.co.jp> wrote:

> As Stephen-san suggested below, could this performance issue be
> improved by setting as below?
> 
>   https://lists.isc.org/pipermail/bind10-dev/2010-November/001573.html
> 
>    PRAGMA synchronous = OFF
>    PRAGMA journal_mode = MEMORY
> 
> >From my quick look, I couldn't find any other codes setting these than
> dhcp-ubench tool.  However this setting might cause a risk of data
> loss.  Otherwise could we consider other tips in the article
> mentioned in the above mail?

As I repeatedly wonder, I suspect this optimization would be marginal
in your case, if it's really an IXFR with a relatively small number of
updates (btw, exactly how many updates in your test scenario?).  If I
understand the above optimization correctly, it's mainly for *bulk*
updates such as updating millions of records at once as mentioned in
the -dev message referenced above.

---
JINMEI, Tatuya
Internet Systems Consortium, Inc.


More information about the bind10-dev mailing list