Maintained by: NLnet Labs

[Unbound-users] unbound-control flush*

Michael Tokarev
Tue Nov 10 10:47:51 CET 2009


A followup to my own post.  It appears to work as expected,
but it definitely works better when pointed to the RIGHT
master with actually up-do-date data ;)

Please excuse me for the false alarm.

/mjt

Michael Tokarev wrote:
> Is it just me or is unbound actually not doing flush*
> operations?
> 
> # unbound-control flush_zone tls.msk.ru
> ok removed 50 rrsets, 83 messages and 0 key entries
> # unbound-control flush_zone tls.msk.ru
> ok removed 50 rrsets, 83 messages and 0 key entries
> # unbound-control flush_zone tls.msk.ru
> ok removed 50 rrsets, 83 messages and 0 key entries
> # unbound-control flush_zone tls.msk.ru
> ok removed 50 rrsets, 83 messages and 0 key entries
> 
> so on each invocation (i did that one right after
> another), it reports that it removed the same amount
> of records.  But actual reports are not being removed,
> since querying it for an RR which changed on master
> (it's a forward zone) returns the same, old data.
> 
> flush and flush_type does not appear to work the
> same way as flush_zone above - both reports a name
> were flushed but in reality it is not.
> 
> Also, unbound-control help says that argument for
> 'flush' is optional, but it returns syntax error if
> no argument is given to it:
> 
> # unbound-control flush
> error cannot parse name
> 
> Restarting the cache clears up the stale data and
> it starts returning new data as it should.  But
> no flush* magic appears to help.
> 
> This is unbound 1.3.4-1, a Debian package.
> 
> Thanks!
> 
> /mjt