locked
Destination Server for Public Folders RRS feed

  • Question

  • Hello,

    I have 3 exchange servers in the organization. One frontend and 2 backends. 2 backends are set to replicate public folders between them. On one of those servers we had to dismount public folder store.

    The effect is that the frontend freezes the queue mails that were sent to the public folders. Frontend necessarily wants to send them to the server on which we had to dismount public folder store.

    Does anybody know how to change destination server for mail sent on public folders?

    Thank you for any advices and suggestions.

    Yours,
    Konrad
    Tuesday, January 13, 2009 12:08 PM

Answers

  • Some problems for you.
    An FE cannot host any mailboxes or PF replicas. Full Stop. Period. No debates. Not possible.

    If the old PF server has a stopped SMTP VSI that's why you've got queued up messages. PF replication is done by SMTP. You will need to get the old server decomissioned or at least remove the PF replica's, wait a while and then stop & remove the PF database.

    Normal email traffic should go over the FE, in and out. That's perfect. But.... unless you've got a very real requirement (and I have never found anyone with one) to send all PF traffic through an FE box on the way to another PF server you wouldn't do it. Sure, if you have a complex infrastructure of several sites each with several servers you're going to send traffic through a local SMTP bridgehead as a funnel on thw way to another server but that's not the same as funneling unecessary traffic through your perimiter FE.

    Sounds like all you need to do is to remove all the replicas from the store and delete it. Kinda. Personally I wouldn't do that because the users on the store need a PF to get free busy information (you're using words FE and BE so I assume E2K3 here) so I would just remove the content replicas and continue to replicate the system folders (OAB, F/B etc.) If you do wish to keep mail flow like it's currently set up it would be ok but honestly I'd change it so that you have the routing group connector in place with the FE as the bridgehead and take the FE address out of the smarthost on the SMTP VSI. (Again, I'm drawing assumptions based on what you've said about what's happening to traffic flow)

    • Proposed as answer by Alan.Gim Friday, January 16, 2009 1:44 AM
    • Marked as answer by Alan.Gim Monday, January 19, 2009 1:01 AM
    Tuesday, January 13, 2009 2:08 PM
  • As Mark said, you shall remove the public folders’ replicas on the old BE, and also the “System Folder” if you want to remove the store on the old BE and it’s the first BE you had created before (Procedure).


    There’re two steps to deliver the messages to public folder, we shall check the attributes in the step one, because it will determine the public store that messages send to. The following is a general process in the step one:

    After server receives messages, it uses categorizer looking to AD, and retrieve public folder object’s “homeMDB” attribute, the attribute points into “msExchOwningPFTreeBL” attribute, which specifies the paths of all stores that make up the public folder hierarchy. Server will send the messages to one of stores based on a special order [Details]

    So, please check the “msExchOwningPFTreeBL” attribute:

    a.       Start->Run->ADSIEdit.msc

    b.      Please find the msExchOwningPFTreeBL path from public folder object’s “homeMDB” attribute

    Location: Expand “Domain” container->” DC=DomainName,DC=Com”->”CN=Microsoft Exchange System Objects”

    c.       Please pick any one of public folder objects, double-click it. In the “Attribute Editor” tab, find “homeMDB” attribute, and double-click the attribute. The value shall look like below:

    CN=Public Folders,CN=Folder Hierarchies,CN=First Administrative Group,CN=Administrative Groups,CN=Company,CN=Microsoft Exchange,CN=Services,CN=Configuration,DC=lab,DC=com

    d.      Please base on that value to find entry “CN=Public Folders

    Example: Expand “Configuration” container->”CN= Configuration,DC=DomainName,DC=Com”->”CN=Services”->”CN=Microsoft Exchange”->”CN=OrgName”->”CN=Administrative Groups”->”CN= AdministrativeGroupsName”->”CN=Folder Hierarchies”

    e.      Please double-click “CN=Public Folders” entry in the right-pane, check “msExchOwningPFTreeBL” in its “Attribute Editor” tab, the values in it shall look like below:

    CN=Public Folder Store (SERVER2),CN=First Storage Group,CN=InformationStore,CN=SERVER2,CN=Servers,CN=First Administrative Group,CN=Administrative Groups,CN=Company,CN=Microsoft Exchange,CN=Services,CN=Configuration,DC=lab,DC=com

    f.        Please keep new BE store’s value unchanged. Select old BE store’s value, and click “Remove” button. Then, removed value will appear in the “Value to add” field, please copy it to notepad for backup.

    g.       Save the change, see if the issue still persists

    Also, you can go to the queue, check stuck messages’ last error info for troubleshooting.

    • Proposed as answer by Alan.Gim Friday, January 16, 2009 1:44 AM
    • Marked as answer by Alan.Gim Monday, January 19, 2009 1:01 AM
    Wednesday, January 14, 2009 9:55 AM

All replies

  • This sounds like you have something set up wrong with your SMTP mail flow because such messages should not be going near the FE. If you have put smarthosts on your SMTP VSIs (on the BE) then this would explain what you're seeing. This is incorrect (generally). If you want mail to go out through the FE you need a Routing Group Connector and you specify the FE as the "Local Bridgehead". This allows PF replication messages to go directly between servers rather than to the FE and back.

    Can you confirm your SMTP configuration on the VSI?
    Tuesday, January 13, 2009 1:10 PM
  • Hi Mark,

    Thank you for an answer!

    Why traffic shouldn't cross by FE? FE is our edge-mail server, connected to Internet. Everyone is allowed to send mail to the public folders (which were on both BE servers). Don't get it. Maybe there is some misunderstanding.

    Moreover, we haven’t got any problems with sending emails to the world, but mails that are send through the Internet, which are mails designated to public folder which should pass through our FE server can't be stored on proper BE server. Old primary public folder server has dismounted public folder store and stopped SMTP Virtual Server. But FE still tries to send mail to the server which generally is down.. We want our FE to use another BE server to store mails designated to public folders. Another BE server's public folder is a replica of old primary public folder server.

    I really hope it is clearer... :)

    Once again thank you very much for your answer and patience!


    Tuesday, January 13, 2009 1:45 PM
  • Some problems for you.
    An FE cannot host any mailboxes or PF replicas. Full Stop. Period. No debates. Not possible.

    If the old PF server has a stopped SMTP VSI that's why you've got queued up messages. PF replication is done by SMTP. You will need to get the old server decomissioned or at least remove the PF replica's, wait a while and then stop & remove the PF database.

    Normal email traffic should go over the FE, in and out. That's perfect. But.... unless you've got a very real requirement (and I have never found anyone with one) to send all PF traffic through an FE box on the way to another PF server you wouldn't do it. Sure, if you have a complex infrastructure of several sites each with several servers you're going to send traffic through a local SMTP bridgehead as a funnel on thw way to another server but that's not the same as funneling unecessary traffic through your perimiter FE.

    Sounds like all you need to do is to remove all the replicas from the store and delete it. Kinda. Personally I wouldn't do that because the users on the store need a PF to get free busy information (you're using words FE and BE so I assume E2K3 here) so I would just remove the content replicas and continue to replicate the system folders (OAB, F/B etc.) If you do wish to keep mail flow like it's currently set up it would be ok but honestly I'd change it so that you have the routing group connector in place with the FE as the bridgehead and take the FE address out of the smarthost on the SMTP VSI. (Again, I'm drawing assumptions based on what you've said about what's happening to traffic flow)

    • Proposed as answer by Alan.Gim Friday, January 16, 2009 1:44 AM
    • Marked as answer by Alan.Gim Monday, January 19, 2009 1:01 AM
    Tuesday, January 13, 2009 2:08 PM
  • As Mark said, you shall remove the public folders’ replicas on the old BE, and also the “System Folder” if you want to remove the store on the old BE and it’s the first BE you had created before (Procedure).


    There’re two steps to deliver the messages to public folder, we shall check the attributes in the step one, because it will determine the public store that messages send to. The following is a general process in the step one:

    After server receives messages, it uses categorizer looking to AD, and retrieve public folder object’s “homeMDB” attribute, the attribute points into “msExchOwningPFTreeBL” attribute, which specifies the paths of all stores that make up the public folder hierarchy. Server will send the messages to one of stores based on a special order [Details]

    So, please check the “msExchOwningPFTreeBL” attribute:

    a.       Start->Run->ADSIEdit.msc

    b.      Please find the msExchOwningPFTreeBL path from public folder object’s “homeMDB” attribute

    Location: Expand “Domain” container->” DC=DomainName,DC=Com”->”CN=Microsoft Exchange System Objects”

    c.       Please pick any one of public folder objects, double-click it. In the “Attribute Editor” tab, find “homeMDB” attribute, and double-click the attribute. The value shall look like below:

    CN=Public Folders,CN=Folder Hierarchies,CN=First Administrative Group,CN=Administrative Groups,CN=Company,CN=Microsoft Exchange,CN=Services,CN=Configuration,DC=lab,DC=com

    d.      Please base on that value to find entry “CN=Public Folders

    Example: Expand “Configuration” container->”CN= Configuration,DC=DomainName,DC=Com”->”CN=Services”->”CN=Microsoft Exchange”->”CN=OrgName”->”CN=Administrative Groups”->”CN= AdministrativeGroupsName”->”CN=Folder Hierarchies”

    e.      Please double-click “CN=Public Folders” entry in the right-pane, check “msExchOwningPFTreeBL” in its “Attribute Editor” tab, the values in it shall look like below:

    CN=Public Folder Store (SERVER2),CN=First Storage Group,CN=InformationStore,CN=SERVER2,CN=Servers,CN=First Administrative Group,CN=Administrative Groups,CN=Company,CN=Microsoft Exchange,CN=Services,CN=Configuration,DC=lab,DC=com

    f.        Please keep new BE store’s value unchanged. Select old BE store’s value, and click “Remove” button. Then, removed value will appear in the “Value to add” field, please copy it to notepad for backup.

    g.       Save the change, see if the issue still persists

    Also, you can go to the queue, check stuck messages’ last error info for troubleshooting.

    • Proposed as answer by Alan.Gim Friday, January 16, 2009 1:44 AM
    • Marked as answer by Alan.Gim Monday, January 19, 2009 1:01 AM
    Wednesday, January 14, 2009 9:55 AM