DLZ data caching -- Please tell me there is *SOME* caching for the db drivers

[Rob Butler]

Simply implement heterogeneous replication as described on this page: http://bind-dlz.sourceforge.net/best_practices.html

It will give you the ease of using Postgres and the performance of BDB.  I recommend the BDBHPT driver.

A "cache" on the other drivers won't be any faster, and it could cause problems.

Rob 

----- Original Message ----
From: Weston Weems <wweems at gmail.com>
To: bind-users at isc.org
Sent: Tuesday, June 26, 2007 5:56:20 PM
Subject: DLZ data caching -- Please tell me there is *SOME* caching for the db drivers

In searching for information on the subject, it would seem I'm not the only
person with this request... but please hear me out...
Current situation, we needed something administed via remote machines (not
through unix, and logins, or putting files here and there, or berkley db)
and for this, postgres+bind dlz has been fantastic. It works EXACTLY as I'd
hoped it would, and with proper replication, I've got to say its easily the
best experience I've had with bind.

The problem, regardless of what everyone says I am not ready and willing to
just laydown without a fight about the subject of at least a small amount of
caching to eliminate a hit per query to postgres database.

Arguments for and against ahve all been heard... and while I dont have a
perfect solution for the problem I'd really like to know why there couldnt
be *SOMETHING*, eg, extra col expected in the postgres DLZ queries for
frequency of looking up against db or SOMETHING.

I've found VERY VERY little documentation on the subject, and was frankly
hoping that some level of caching at the dlz level has been implemented and
I am just missing it.

Thanks for any advice

Weston Weems









 
____________________________________________________________________________________
Need Mail bonding?
Go to the Yahoo! Mail Q&A for great tips from Yahoo! Answers users.
http://answers.yahoo.com/dir/?link=list&sid=396546091