Using named_dump.db to preserve cache across reboots
Mr Toad
I'm running BIND 8.2.3 on Debian Linux 2.2 and would like to be able
to preserve the cached RRs across reboots.
I know that sending SIGINT to the named process causes bind to dump
its database to "named_dump.db", and I wonder if this can be used to
preserve the cached records so they could be reloaded after a reboot.
Is this a supported or reasonable thing to do, or is the named_dump.db
dumpfile not suitable for this?
If it is suitable, how should I get bind to load it up at startup?
Should I enter it into the named.conf file as a zone file?
I'm using the nameserver both as a primary for my local forward and
reverse zones and also as a caching nameserver for my Internet
queries.
Apologies if this is an FAQ question - it seems like it is, but I
can't see it answered anywhere. If it is described somewhere, please
let me know which FM I should R.
Mr Toad
Barry Margolin
Mr Toad <rat_lo...@yahoo.co.uk> wrote:
>
>I'm running BIND 8.2.3 on Debian Linux 2.2 and would like to be able
>to preserve the cached RRs across reboots.
>
>I know that sending SIGINT to the named process causes bind to dump
>its database to "named_dump.db", and I wonder if this can be used to
>preserve the cached records so they could be reloaded after a reboot.
>
>Is this a supported or reasonable thing to do, or is the named_dump.db
>dumpfile not suitable for this?
It's not suitable. BIND has no built-in provision for preserving the
cache. named_dump.db is for diagnosing problems (e.g. finding out who
poisoned your cache).
--
Barry Margolin, bar...@genuity.net
Genuity, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.
Mr Toad
On 8 Mar 2002 09:03:01 -0800, Barry Margolin <bar...@genuity.net>
wrote:
>It's not suitable. BIND has no built-in provision for preserving the
>cache. named_dump.db is for diagnosing problems (e.g. finding out who
>poisoned your cache).
That's a shame - I'd hoped that there would be some way to preserve
the cache.
Does anyone know if this feature is available for BIND 9, or planned
for the future?
Mr Toad
Jim Reid
>> It's not suitable. BIND has no built-in provision for
>> preserving the cache. named_dump.db is for diagnosing problems
>> (e.g. finding out who poisoned your cache).
Toad> That's a shame - I'd hoped that there would be some way to
Toad> preserve the cache.
Why? What's the point of preserving a cache that is by definition now
out of date? Would you expect the weather forecast to tell you what
the weather was like 2 weeks ago? And if it did, what use would that
be to anyone?
Toad> Does anyone know if this feature is available for BIND 9, or
Toad> planned for the future?
It's not. I suppose if you chuck enough money at ISC they might
consider implementing this feature if you can provide convincing
arguments for it. Quoting RFC1925 "With sufficient thrust, pigs fly
just fine."
Mr Toad
On 11 Mar 2002 09:16:28 -0800, Jim Reid <j...@rfc1035.com> wrote:
>Why? What's the point of preserving a cache that is by definition now
>out of date? Would you expect the weather forecast to tell you what
>the weather was like 2 weeks ago? And if it did, what use would that
>be to anyone?
The reason for wanting to preserve the cache is because, over time, my
nameserver builds up a rich cache which speeds up lookups for
commonly-requested names. Now if I reboot the machine, that cache is
lost and will take time to grow back. I would prefer to be able to
preserve the cache across reboots so that it is not lost.
I can't understand why you say that it's like asking for last week's
weather forecast. Surely the nameserver cache is just as valid after
a reboot as it was a couple of minutes before?
Mr Toad
p...@icke-reklam.ipsec.nu
Mr Toad <rat_lo...@yahoo.co.uk> wrote:
Sound to me that avoiding reboots would cure the problems.
> Mr Toad
--
Peter Håkanson
IPSec Sverige (At the Riverside of Gothenburg, home of Volvo)
Sorry about my e-mail address, but i'm trying to keep spam out.
Remove "icke-reklam" and it works.
Mark Damrose
"Mr Toad" <rat_lo...@yahoo.co.uk> wrote in message
news:a6lcj9$d...@pub3.rc.vix.com...
I don't see the big deal. Records in the cache expire and are looked up
again all the time. For the popular names, they will be looked up again
quickly, then they will be in your cache.
However, more and more of the popular services use load balancers. They
have very short TTL on their records, so that adjustments are seen quickly.
This makes your cache less usefull. When you have a TTL of 5 minutes and a
reboot takes 2-3, it's not worth reloading that record.
If you really want this cache to be re-built, you could always write a
script to parse the named_dump.db file and lookup everything in it.
>
> Mr Toad
>
>
Barry Margolin
Mr Toad <rat_lo...@yahoo.co.uk> wrote:
>The reason for wanting to preserve the cache is because, over time, my
>nameserver builds up a rich cache which speeds up lookups for
>commonly-requested names. Now if I reboot the machine, that cache is
>lost and will take time to grow back. I would prefer to be able to
>preserve the cache across reboots so that it is not lost.
The time it would take to read in a huge cache file (our dump files are
usually hundreds of megabytes) and recreate all the cache entries would
probably negate the value of this. If it adds 1 minute to the restart
time, that's time that could have been spent performing hundreds or
thousands of recursive queries that would have served the users *and*
refilled the cache with many of the important entries.
Jim Reid
Toad> The reason for wanting to preserve the cache is because,
Toad> over time, my nameserver builds up a rich cache which speeds
Toad> up lookups for commonly-requested names. Now if I reboot
Toad> the machine, that cache is lost and will take time to grow
Toad> back. I would prefer to be able to preserve the cache
Toad> across reboots so that it is not lost.
Well if you think the server's cache is *really* that precious, stop
rebooting it!
Bob
Home made scripts are always fun...
Mr Toad
On 12 Mar 2002 10:37:44 -0800, Jim Reid <j...@rfc1035.com> wrote:
>Well if you think the server's cache is *really* that precious, stop
>rebooting it!
Thanks to everyone who chipped in on this thread. I'm now convinced
that I'm barking up the wrong tree regarding cache preservation.
Mr Toad