How often is it suggested to restart Bind8 ?

[Lawrence Chan]

Barry Margolin wrote:

> In article <4.2.0.58.20000209182939.0441dbd0 at pop3.mailbox.co.uk>,
> Andy Spiers  <andy at centralnic.com> wrote:
> >At 11:36 AM 2/9/00 -0500, Steve Snyder wrote:
> >>2. What happens when the max size is reached?  Are older data cleaned
> >>out when new data arrives?
> >
> >I have always assumed that this is the case but I have never seen any
> >conclusive documentation on the method used to select cached RRs to discard.
>
> As I wrote in my earlier message, your assumption is wrong.
>
> named uses the limit XXXsize options as parameters in the Unix ulimit()
> system call, which tells the OS to refuse requests to grow that particular
> memory segment beyond that size.  When BIND gets an error due to one of
> these limits, it crashes.
>
> You could have a cron job that checks whether named is running and restarts
> it if necessary, or you could run it in a loop in a script:
>
> while /bin/true; do
>   named -d
> done
>
> The -d option is necessary to prevent it from backgrounding itself (which
> would cause the loop to try spawning named continuously).
>
> --
> Barry Margolin, barmar at bbnplanet.com
> GTE Internetworking, Powered by BBN, Burlington, 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.

Is there a default value for such cache size set when installing Bind and if so,
how large?  Surely, this sounds like a disaster waiting to happen if one has
limited memory and without some sort of cron job.  Is this problem applicable to
all Bind versions or is it just to Bind 8 onward?

Lawrence Chan