Storing Duplicate Copies of Quarantine

[Glenn Steen]

On 17/01/2008, Brian O'Keeffe <brian.okeeffe at kepak.com> wrote:
> Glenn Steen wrote:
> > On 17/01/2008, Brian O'Keeffe <brian.okeeffe at kepak.com> wrote:
> >
> >> Hi all,
> >>
> >> I'm running two mail gateways with MailScanner and MailWatch, I've
> >> trying to give users access to release their quarantined mails using
> >> MailWatch's quarantine report but want to give users a single point of
> >> contact rather than getting reports from both gateways and releasing
> >> mails from both.
> >> I've set up NFS shares so that the gateways can write to each other and
> >> set up Master-Master replication on the MySql databases on both. Where
> >> I'm having a problem is getting MailScanner to write Quarantined files
> >> to the quarantine folder on the box that its running on and also to the
> >> quarantine on the other gateway.  I've got it writing to the archive on
> >> both machines but just can't get the quarantine to do the same. Is there
> >> some way I can write a ruleset to achieve this?
> >>
> >> Any assistance, ideas or comments would be very gratefully received.
> >>
> >> Regards
> >>
> >> Brian
> >>
> >>
> > Why go through all this, when MailWatch can log to one DB and release
> > (through XML-RPC) from both?
> >
> > Cheers
> >
> Thanks Glenn but I'm trying to build in redundancy as well, so if one
> gateway drops then the other still has everything necessary and at most
> all i need do is activate the MailWatch Quarantine report. Am I
> misreading the MailWatch documentation or can I use the XML-RPC to do this.
>
Hi Brian,

Hm, nah I don'tmthink you are, not wrt redundacy.
Thing is you'll have to take XML-RPC and how this all is implemented
in MW into account anyway, since the release functions look at where
the quarantined item is stored (on which machine). Not that big a
deal, just something that need be handled, if one (like you) plan to
have the machines be completely redundant.
The XML-RPC stuff simply place a call to do the action on the "other
host" as needed, so ... not exactly what you are trying to do.

What I meant was that your approach is rather more complex than using
the simple thing that is already present, for data that mostly is very
... "losable":-):-).

Cheers
-- 
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se