Some one please take a look at my steps and verify. (Full Version)

All Forums >> [Microsoft Exchange 2007] >> Installation



Message

frank.merino -> Some one please take a look at my steps and verify. (19.Nov.2007 2:31:47 PM)

Hello Everyone,

I have a 2003 Exchange Server on a Win2k3 box and i'm looking to movie over to an Exchange 2007 with CCR.

So here is what i have and what i think i'm suppose to do please someone correct me if need be.

1. setup one box with HT/CS leave the mailboxes on 2k3
2. setup 2 boxes with CCR and the mailbox role
3. move mailboxs and public folder over
4. decomission 2k3

Please let me know if this is correct. I'm not sure on how to do this but i'll figure it out if this is correct. Thanks in advance.

Frank



a.grogan -> RE: Some one please take a look at my steps and verify. (19.Nov.2007 2:50:10 PM)

Hiya Frank,

A couple of things, that may help;

1. Run the Exchange 2007 readiness check in your 2003 environment first - you need to ensure that there are no instances of Exchange 5.5, no instances of SRS, all ADC's are removed and the org is running in native mode.

2. Ensure that DC connectivity is ok by using the DCDIAG test
3. From the Exchange 2007 CD on your x64 server prepare the domain using the setup.com /PrepareAD switch (this will take care of legacy permissions, universal groups, legacy groups, provision for the 2003 RUS, import the schema and places the configuration in AD)

4. The first roles in should be the CAS (which you would combine with the HT role) - during setup you will be asked to provide a legacyroutinggroup server (which should be the routing group master in your Exchange 2003 installation)

5. Remember that with CCR there should only be 1 instance of public folders in the entire organisation - therefore, you Exchange 2003 public folder will continue to work, and they will replicate over to your 2007 box - you should aim to de-commission the 2003 PF as quickly as possible as the PF on the CCR will not fail-over properly.

6. When you install the HT role, and being to send messages between server - if nothing appears to be happening, do a search for "back pressure".

There are loads more things to look out for, but that should get you started - and remember BACKUPS - DC's and Exchange servers.

Cheers

A



frank.merino -> RE: Some one please take a look at my steps and verify. (19.Nov.2007 4:50:17 PM)

hello a.grogan,

Thank you for the quick reply. when you talked about getting rid of the PF from 2003 will it replicate once the 2k3 is gone? i have one storage group in 2k3 right now and place to have 1 storage group per database in 2k7. is that what you would recommend?

Thanks again for the help and i will let you know how it goes.



a.grogan -> RE: Some one please take a look at my steps and verify. (20.Nov.2007 1:44:40 PM)

Hiya Frank, what you need to be mindful of is, although the Exchange 2003 PF will replicate to the CCR PF database, if the CCR fails over due to an issue (often referred to as lossy fail-over) the PF store will not mount on the passive node of the CCR cluster.

Therefore it is advisable that you get all content from your Exchange 2003 databases over to the CCR PF store as quickly as possible, decommission the 2003 PF store and re-point all of you clients at the CCR PF and make sure that remains the only PF instance in the Organisation.

Cheers

A



frank.merino -> RE: Some one please take a look at my steps and verify. (20.Nov.2007 3:56:05 PM)

So will the PF be ok as long as the exchange ccr doesn't fail over? Is shouldn't failover unless it actually does fail correct?

Why is there no guide about how to do this?

Frank



Exchange.Specialist -> RE: Some one please take a look at my steps and verify. (25.Nov.2007 2:22:02 PM)

Sorry, but i dont agree upon that mate. If Pf was such a pain with CCR. you would have seen a report from MS about not to work with PF on CCR. Yes working with more than 1 PF DB (on E2k / E2k3 box) along with PF DB on CCR was earlier not supported however with revised testing MS now completley supports it.

So, coming back to the point - PF replication does take about 48-72 hrs to complete. My bet would be to check ESM for item count for entire PF replication between E2k3 and E2k7 (again understand im talkn abt entire content replication and not simply hierarchy) to complete.

Post replication and move mailboxes, upgrade all your address list to E2k7. Shut the E2k3 box for 2 days to check the entire functionality of E2k7. Then take a stand about removing E2k3 completely.

All the best..............



frank.merino -> RE: Some one please take a look at my steps and verify. (26.Nov.2007 12:12:29 PM)

Hello,

I have been following this HOWTO: http://www.msexchange.org/tutorials/Transitioning-Exchange-2000-2003-Exchange-Server-2007-Part2.html and i'm about to install exchange HT and CAS right now. will i lose connection to Exchange 2003 mailboxes while the install and the routing of mail goes through exchange 2007? 

Thanks,
Frank



a.grogan -> RE: Some one please take a look at my steps and verify. (26.Nov.2007 12:48:45 PM)

Specialist, I would be interested to see a link to this - from my point of view the documentation still states that if a lossy fail over occurs on a CCR cluster which is installed in an environment with more than one PF database (for example Exchange 2003 or other mailbox servers which are 2007 based which contain a PF database) when attempting to mount the PF database on the CCR  passive node the resource will fail.
This is in contrast to using the move-clusteredMailbox command on the CCR which will move the resource and mount the database correctly - however the keyword here is "lossy" e.g. and error occurs.
I have not seen this corrected as of yet, nor the recommendation from Microsoft where they request migrating public folder content as a high priority if you have more than one public folder database lowered to a do it in your own time - it still states do it ASAP.
My information is as per http://technet.microsoft.com/en-us/library/bb123996.aspx
Also I never said that it was not supported I was explaining to frank that there can be issues when operating more than one PF database which is replicating in an organisation with a CCR cluster.

Frank in answer to your question - no you shouldn't have any issues.

Cheers



frank.merino -> RE: Some one please take a look at my steps and verify. (26.Nov.2007 1:21:59 PM)

ok i have 07 HT and CAS installed i don't know yet whether it's working properly or not. no errors during the install and external/internal mail is still flowing so that' s good. I just don't know if it is going through my 07 server. Is there a way to check?

Also assuming all is good the next set would be to setup the Cluster, Test the cluster, install CCR and then start moving mailboxes and the PF correct?


Thanks,
Frank 



a.grogan -> RE: Some one please take a look at my steps and verify. (26.Nov.2007 4:05:59 PM)

Hiya Frank, yep you should be ok to proceed - however, be mindful of ensuring the HT role is adequately resourced for the volume of mail that it will have to deal with - the following article on "back pressure"will explain more http://technet.microsoft.com/en-us/library/bb201658.aspx

In terms of the CAS we can go into further detail with this when you have your cluster configured and setup.

Cheers

A



frank.merino -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 9:05:23 AM)

Hello Andy,

I wanted to know something. I'm about to start this part 2 of 3 for the CCR setup (http://www.msexchange.org/tutorials/Installing-Configuring-Testing-Exchange-2007-Cluster-Continuous-Replication-Based-Mailbox-Server-Part2.html) and i was wondering once this is done when do the clients see a lost of connections to the Exchange server?

Would i be able to bring up the CCR and then the lost of connection would happen when i move the physical mailbox?


Thanks again for all your help so far and in the future.
Frank



a.grogan -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 11:05:45 AM)

Hiya Frank, by adding the CCR to the installation you should not at this point see any mailbox disconnections.
Mailbox moves generally should be done when the user is not in the mailbox - but the move process sets a flag which lets the system know that the mailbox is being moved therefore messages to it should be queued.

What I would suggest is to move a test mailbox over -or- you own prior to anyone else in your organisation - this will give you the chance to ensure that the CCR, CAS, and HT are functioning correctly and you can gain access to OWA, use mobile devices and send and recieve mail.

Cheers
A



frank.merino -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 11:15:50 AM)

Thanks Andy, i will let you know if i run into any issues along the way.



frank.merino -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 1:12:22 PM)

Hey Andy i just got this error on the first node in the CCR

Summary: 3 item(s). 2 succeeded, 1 failed.
Elapsed time: 00:13:32

Copy Exchange Files
Completed
Elapsed Time: 00:01:06

Mailbox Role
Completed
Elapsed Time: 00:10:51

Clustered Mailbox Server
Failed
Error:
Cluster Common Failure Exception : The specified node does not support a resource of this type.  This may be due to version inconsistencies or due to the absence of the resource DLL on this node. (Exception from HRESULT: 0x800713D7)
The specified node does not support a resource of this type.  This may be due to version inconsistencies or due to the absence of the resource DLL on this node. (Exception from HRESULT: 0x800713D7)
Warning:
An unexpected error has occurred and a Watson dump is being generated: Cluster Common Failure Exception : The specified node does not support a resource of this type.  This may be due to version inconsistencies or due to the absence of the resource DLL on this node. (Exception from HRESULT: 0x800713D7)

Elapsed Time: 00:01:34


I'm looking it up now but just in case you have seen it before.

Frank



a.grogan -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 1:51:45 PM)

Hiya Frank,

Just going to run through a couple of things;

1. Did the MNS clustering installation go ok?
2. What is the configuration of the cluster (e.g. MNS quorum location and how are the disks attached)
3. Interface configuration
4. Was the Exchange prerequisite software installed ok (MMC 3.0, .NET 2.0 + fix, IIS with No SMTP or NNTP)?
5. Was the clustering account a member of the local admins group on the Node
6. Did the account that you used to install Exchange have the correct permissions on the local machine (e.g. admin) and the permissions to update the Directory.

Cna you have a look and let us know,

Cheers

A



frank.merino -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 2:04:31 PM)

1. Did the MNS clustering installation go ok?
- Yes no Errors
2. What is the configuration of the cluster (e.g. MNS quorum location and how are the disks attached)
- MNS is located on the HT and CAS using a Share Folder. i see that some items have been written to the share (i can attach if you would like)
3. Interface configuration
- Not sure i understand this q?
4. Was the Exchange prerequisite software installed ok (MMC 3.0, .NET 2.0 + fix, IIS with No SMTP or NNTP)?
- MMC ok (server is 2k3 x64 R2 SP2) .net 2.0 installed and server fully patched if you know the KBs i need i will double check., IIS is install, SMTP and NNTP is not installed
5. Was the clustering account a member of the local admins group on the Node
- Yes, but not the HT and CAS Server (do i need to make him an admin)
6. Did the account that you used to install Exchange have the correct permissions on the local machine (e.g. admin) and the permissions to update the Directory.
- Account is Domain Admins, Enterprise Admins, Exchange Organization Administrators, Schema Admins

Please let me know

Thanks,
Frank



a.grogan -> RE: Some one please take a look at my steps and verify. (27.Nov.2007 2:09:05 PM)

Hiya chap, ok in the following location;
c:\ExchangeSetupLogs - you will find a number of files, do you mind zipping them up and mailing them to the MSN address in my profile.

Cheers

A



frank.merino -> RE: Some one please take a look at my steps and verify. (28.Nov.2007 9:48:16 AM)

Hey Andy,

I was able to get the cluster to come back up without any issues so i'm ready to start the exchange install again let me know if you want me to send you anything before i start.


Frank



frank.merino -> RE: Some one please take a look at my steps and verify. (30.Nov.2007 8:23:59 AM)

Thank You so much Andy for the help,

It looks like there was a failed install on the 2nd node that did not get cleaned out fully and we had to remove the Network Monitor Driver from the nics to improve the cluster failover speed. I also update the NIC drivers which were apparently updated this month.

After that tried again and i have a working cluster and the CCR is in and tested.

So now the next step is to start moving mailboxes over correct? Obviously move test accounts first



a.grogan -> RE: Some one please take a look at my steps and verify. (30.Nov.2007 11:31:59 AM)

Yep We had some fun working that out.
Essentially Microsoft NetMon installs a packet monitor on each interface on the node - after unticking this cluster functionality picked up and therfore stopped failing over during Exchange Setup (which was causing the error message that Frank was seeing).
Coupled with an incomplete install on the passive node (which we removed by hand) - we got there.
Anyway as you have said if you grab a few test MB's over onto the cluster - see if you can hit them with the via the CAS and send mail to them.

Cheers

A



Page: [1] 2   next >   >>