Hi, We are migrating from Exchange 2003 to 2010, Exchange 2003 on 2003 R2 Sp2 to 2008R2 with 2010 Sp1, about 800 Mailboxes. The mailboxes are on the new server, working fine. The Problem: We have about 400 GB of Public Folders, right now we are in the middle of the first synchronization/Replication. about 200 GB of the store is synced to the new server After about 30 Minutes of rebooting the 2003 server or restarting the services, the queue grows from about 500 items to >10000, somtimes more than 30000. I can see in system event log: Event Type: Error Event Source: smtpsvc Event Category: None Event ID: 402 Date: 22.03.2011 Time: 22:46:44 User: N/A Computer: 2003 Exchange Description: Virtual Server 1: IPADDRESS OF 2003 EXCHANGE maximum number of connections has been reached. Connection being closed. For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp. In Application log many of these show up: Event Type: Warning Event Source: MSExchangeTransport Event Category: SMTP Protocol Event ID: 7002 Date: 23.03.2011 Time: 00:36:29 User: N/A Computer: 2003 EXCHANGE Description: This is an SMTP protocol warning log for virtual server ID 1, connection #891. The remote host "2010 EXCHANGE", responded to the SMTP command "quit" with "421 4.4.1 Connection timed out ". The full command sent was "QUIT ". This may cause the connection to fail. For more information, click http://www.microsoft.com/contentredirect.asp. and Event Type: Warning Event Source: MSExchangeTransport Event Category: Exchange Store Driver Event ID: 327 Date: 23.03.2011 Time: 00:14:58 User: N/A Computer: 2003 EXCHANGE Description: The following call : EcOpenItemForLocalDelivery to the store failed. Error code : -2147221233. MDB : . FID : . MID : . File : . For more information, click http://www.microsoft.com/contentredirect.asp. Do you have any idea why this happens and how to solve it? Thanks in advance Markus

Have you increased the replication message size for those public folder stores? The default sizes are way too small. The small message sizes are causing lots of replication messages to be created (not that there won't still be lots of them!). I'd probably bump that size up to 4MB (or more) on both sides on all PF databases. The Exchange 2003 server i trying to send the replication messages as fast as possible to the Exchange 2010 server, but the 2010 server is throttling the number of inbound connections so it's not going to be overwhelmed with connections (especially from a single source), or with a flood of messages from a singel source. You can modify that with the Set-ReceiveConnector cmdlet and the -MaxInboundConnection, -MaxInboundConnectionPercentagePerSource, -MaxInboundConnectionPerSource, and -MessageRateLimit parameters.--- Rich Matheisen MCSE+I, Exchange MVP

Any update?Gulab | MCTS-MCITP Messaging: 2010 | MCTS-MCITP Messaging: 2007 | MCC 2011 | Skype: Gulab.Mallah

Dear Rich, thank you,i will change the settings and keep you informed. Markus

I have made the following changes: I increased the size to 4096 kb on both servers I increased the -MaxInboundConnections on the 2010 Server from 5000 to 50000 (the other parameters are 100 percent and unlimited) . restarted exchange transport and mail submission on the 2010, made an iisreset on 2003 and restarted the mailbox store. Since the change of the maxinboundconnections the 2010 gives the following error in application log about once a second: Log Name: Application Source: MSExchange Store Driver Date: 3/23/2011 12:27:43 PM Event ID: 1020 Task Category: MSExchangeStoreDriver Level: Error Keywords: Classic User: N/A Computer: EX2010.DOMAIN Description: The store driver couldn't deliver the public folder replication message "Folder Content (Ex2003-IS@domain)" because the following error occurred: The process failed to get the correct properties.. Event Xml: <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event"> <System> <Provider Name="MSExchange Store Driver" /> <EventID Qualifiers="49156">1020</EventID> <Level>2</Level> <Task>1</Task> <Keywords>0x80000000000000</Keywords> <TimeCreated SystemTime="2011-03-23T11:27:43.000000000Z" /> <EventRecordID>390199</EventRecordID> <Channel>Application</Channel> <Computer>Ex2010</Computer> <Security /> </System> <EventData> <Data>"Folder Content (EX2003-IS@domain)"</Data> <Data>The process failed to get the correct properties.</Data> </EventData> </Event> Anything else is unchanged

I have made the following changes: I increased the size to 4096 kb on both servers I increased the -MaxInboundConnections on the 2010 Server from 5000 to 50000 (the other parameters are 100 percent and unlimited) . restarted exchange transport and mail submission on the 2010, made an iisreset on 2003 and restarted the mailbox store. Since the change of the maxinboundconnections the 2010 gives the following error in application log about once a second: Log Name: Application Source: MSExchange Store Driver Date: 3/23/2011 12:27:43 PM Event ID: 1020 Task Category: MSExchangeStoreDriver Level: Error Keywords: Classic User: N/A Computer: EX2010.DOMAIN Description: The store driver couldn't deliver the public folder replication message "Folder Content (Ex2003-IS@domain)" because the following error occurred: The process failed to get the correct properties.. Event Xml: <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event"> <System> <Provider Name="MSExchange Store Driver" /> <EventID Qualifiers="49156">1020</EventID> <Level>2</Level> <Task>1</Task> <Keywords>0x80000000000000</Keywords> <TimeCreated SystemTime="2011-03-23T11:27:43.000000000Z" /> <EventRecordID>390199</EventRecordID> <Channel>Application</Channel> <Computer>Ex2010</Computer> <Security /> </System> <EventData> <Data>"Folder Content (EX2003-IS@domain)"</Data> <Data>The process failed to get the correct properties.</Data> </EventData> </Event> Everything else is unchanged

I don't think raising the maximum number of connections to 50,000 is wise. IPv4 limits the machine to using no more than 64K ports and each connection is going to use one port. That doesn't leave a lot of ports available in the 1K - 64K range. However, I don't think that's your problem because I don't think E2K3 can support that many concurrent connections. Back that value back down to something reasonable. The problem with the message content (the correct property) may be evident if you enable pipeline and content conversion tracing. Rather than turning up replication on ALL public folders, why not start out with some small subset of folders and make sure that works first?--- Rich Matheisen MCSE+I, Exchange MVP

It is back to 5000, but this has not been the problem. The initial replication is done, I have rebooted both servers, no more errors up to now. It has not been my idea to start with all of them and I have not turned up replication on all of them, I have just been ask to help with the troubleshooting. Thank you for your help and one question. Is the default replication interval of 15 minutes recommended for PF replication of that size or should I increase that value? Markus

The replication interval is just how often the PF store will send any updates. If there have been no changes then the server will send no replication messages for the folder. You can increase it, but it will mean that the notification of changes in a folder won't be sent too replication partners as often. If you still have people using public folder for free/busy then anything they change in their calendar won't be sent to the other public folder store as often. Since that can lead to double-booking of time you should think about that before you increase the interval.--- Rich Matheisen MCSE+I, Exchange MVP

I have changed it and replication seems to work now. The errors and problems were caused by the amount of folders that had to be replicated at the same time. As soon as the initial replication was finished - no more errors. Thank you for your help. Markus