Re: Exchange and AD Redundancy
From: BP (inquire_at_glipit.com)
Date: 05/08/04
- Previous message: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- In reply to: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- Next in thread: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- Reply: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- Messages sorted by: [ date ] [ thread ]
Date: Sat, 8 May 2004 09:32:40 -0400
Yes,
Exchange shares the schema with AD so I think your last comment
was somewhat harsh. How do you know the delay Andy spoke
of was not from KCC trying to recover the lost intersite connection
before it would allow Exchange to continue access on the schema.
Without us seeing the event logs it is anyone's best guess.
"Boris Lokhvitsky" <msexpert@comcast.net> wrote in message
news:eGDu1XJNEHA.3052@TK2MSFTNGP12.phx.gbl...
> Hello,
> Thanks for the link to the article. The phrase I liked most in there, is the
> last phrase: "Preferred Bridgehead servers should not be defined unless
> absolutely necessary." :) which you actually already mentioned in your
> earlier post.
> Well, as can be seen from there too, actually ANY domain controller acts as
> a KCC. You can designate some DC to be a preferred bridgehead, but if you
> don't, any domain controller can be used for intersite replication. Which is
> not bad at all, in my understanding.
> Also, KCC is NOT one of the FSMO roles. So no surprise that it is not
> unique, and multiple domain controllers hold this role.
> FSMO roles act independently and differently from KCC, and they do not fail
> over automatically... if your FSMO role owner (no matter which role)
> crashed, you are in trouble. There are multiple KB articles explaining how
> to seize FSMO roles in this case.
> In any case, all this KCC and bridgehead stuff refers to intersite AD
> replication and has nothing to do with Exchange at all.
>
> Regards,
> Boris
>
>
> "BP" <inquire@glipit.com> wrote in message
> news:OHnjYEJNEHA.268@TK2MSFTNGP11.phx.gbl...
> > Although somewhat contradicting within the article top text to bottom text
> > it must be assumed in article more than two servers exist within site that
> > could hold the KCC role?
> >
> > Article - 271997 Not the original looks updated from 2001
> >
> > The KCC chooses bridgehead servers in the following manner:
> >
> > As the KCC constructs the inter-site topology for each directory
> partition, the servers in
> > each site are evaluated to see if they can be bridgehead servers.
> Preferred bridgehead
> > servers are chosen first. If you have not configured a preferred
> bridgehead server, all
> > servers in the site that host the directory partition, and that can
> communicate over a
> > given transport, can become bridgehead servers. In either case, the first
> domain
> > controller that meets the requirements becomes the bridgehead server. The
> domain
> > controllers are sorted in ascending globally unique class identifier
> (GUID) order. Each
> > preferred bridgehead server that exists in the same site that is
> configured for the same
> > transport is considered equal.
> >
> >
> > This must have been what I read and confused standby FSMO within same site
> and checking a
> > default
> > Bridgehead for IP/Smtp transports KCC but then again what exactly do they
> mean by standby
> > FSMO
> >
> > Article - 223346
> > FSMO Availability and Placement
> > Windows 2000 performs the initial placement of roles on domain
> controllers. This placement
> > is often correct for directories with few domain controllers. In a
> directory with many
> > domain controllers the default placement is unlikely to be the best match
> to your network.
> >
> > On a per-domain basis, select local primary and standby FSMO domain
> controllers in case a
> > failure occurs on the primary FSMO owner. Additionally, you may want to
> select off-site
> > standby owners in the event of a site-specific disaster scenario. Consider
> the following
> > in your selection criteria:
> > a.. If a domain has only one domain controller, that domain controller
> holds all the
> > per-domain roles.
> > b.. If a domain has more than one domain controller, use Active
> Directory Sites and Serv
> > ices Manager to select direct replication partners with persistent,
> "well-connected"
> > links.
> > c.. The standby server may be in the same site as the primary FSMO
> server for faster
> > replication convergence consistency over a large group of computers, or in
> a remote site
> > in the event of a site-specific disaster at the primary location.
> > d.. Where the standby domain controller is in a remote site, ensure that
> the connection
> > is configured for continuous replication over a persistent link.
> >
> >
> > "Boris Lokhvitsky" <msexpert@comcast.net> wrote in message
> > news:u1JiJ9ENEHA.3636@TK2MSFTNGP09.phx.gbl...
> > > I do not think any DC can automatically "take over" the FSMO roles
> unless
> > > you manually demote the original roles owner. That is why they are "SM",
> > > after all. But if you can find the link to an article you mentioned I'd
> > > really appreciate it.
> > >
> > > Thanks,
> > > Boris
> > >
> > >
> > > "BP" <inquire@glipit.com> wrote in message
> > > news:uS3DXG%23MEHA.2532@TK2MSFTNGP10.phx.gbl...
> > > > Microsoft printed an article about not assigning
> > > > any server as preferred bridgehead in AD sites for smtp
> > > > or IP transports where multiple DC exist so other DC's still
> > > > operational with in the site can take over the functions
> > > > like knowledge consistency and believe FSMO roles.
> > > > I looked in my article archives but seem to have lost it.
> > > >
> > > > "Andy" <anonymous@discussions.microsoft.com> wrote in message
> > > > news:950501c43381$51b3eb80$a501280a@phx.gbl...
> > > > > Hello there,
> > > > >
> > > > > I have a senerio I was hoping someone could shed some
> > > > > light on.
> > > > >
> > > > > I have two domain controllers, the 1st DC contains all of
> > > > > the FSMO roles and a global catalog. I have one Exchange
> > > > > 2000 server. When I shut down the 1st DC, Outlook clients
> > > > > have trouble using or opening outlook and communicating
> > > > > with the email server.
> > > > > To resolve this I added a second global catalog to the
> > > > > second DC. Now when I shut down the first clients still
> > > > > indicate problems. My workstation seems to have trouble
> > > > > connecting for a couple of minutes but then I was able to
> > > > > connect. This did not work as seamless as I had been
> > > > > hoping, however I only had the DC down for about 5
> > > > > minutes, does it take longer for the second DC to pick up
> > > > > the slack?
> > > > >
> > > > > Thanks for your help
> > > > > Andy
> > > >
> > > >
> > >
> > >
> >
> >
>
>
- Previous message: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- In reply to: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- Next in thread: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- Reply: Boris Lokhvitsky: "Re: Exchange and AD Redundancy"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|