[bind10-dev] it takes long time to IXFR

Yoshitaka Aharen aharen at jprs.co.jp
Wed Mar 27 04:17:30 UTC 2013


Hello,

Yes, we can apply a patch to the experimental server if it is needed.
However, I must hide some part of logs as it may contain sensitive
information.
I haven't tried yet but the situation could be reproduced with creating
a zone with some millions of records, I guess.

Thanks,

On Fri, 22 Mar 2013 10:59:48 -0700
JINMEI Tatuya / 神明達哉 <jinmei at isc.org> wrote:

> At Fri, 22 Mar 2013 15:05:52 +0900,
> Yoshitaka Aharen <aharen at jprs.co.jp> wrote:
> 
> > I've configured severity: DEBUG, debuglevel: 40 (DBGLVL_TRACE_BASIC).
> > It shows it takes very long time between DATASRC_SQLITE_CONNOPEN and
> > DATASRC_SQLITE_DROPCONN, which seems to be doing actual database update.
> 
> Hmm, we don't seem to provide detailed trace logs within the update
> transaction...is it possible to apply a patch to that experimental
> server to produce more details logs, e.g, for each INSERT/DELETE SQL
> statements, if we provide such a quick hack patch?  Then we might be
> able to identify the most significant bottleneck.
> 
> ---
> JINMEI, Tatuya
> Internet Systems Consortium, Inc.
> 

-- 
Yoshitaka Aharen <aharen at jprs.co.jp>
Japan Registry Services Co., Ltd.




More information about the bind10-dev mailing list