BNettles73
asked on
Messages awaiting directory lookup - Email Queuing
Messages are queuing in the "Messages awaiting directory lookup" queue.
General Configuration information:
Exchange 2003 SP 1 Server A
Internet Mail Connector 1 configured as a Smart Host to DMZ Server A (IIS 6.0 SMTP Gateway) with a cost of (1)
(4) 2.0 GHz Procs - 4GB Memory - Tuned for Large Memory
GB Ethernet - second adapter disabled
Exchange 2003 SP1 Server B
Internet Mail Connector 2 configured as a Smart Host to DMZ Server B (IIS 6.0 SMTP Gateway) with a cost of (1)
(4) 2.0 GHz Procs - 4GB Memory - Tuned for Large Memory
GB Ethernet - second adapter disabled
Exchange 2003 SP1 Server C
Internet Mail Connector 1 configured as a Smart Host to DMZ Server A (IIS 6.0 SMTP Gateway) with a cost of (10) ** Server used for Application Server relay
(4) 2.0 GHz Procs - 4GB Memory - Tuned for Large Memory
GB Ethernet - second adapter disabled
8 Procs on the E2K3 SP1 MBX Cluster A / 4 Procs on the (2) E2K3 Public Folder Servers for a total of 28 E2K3 Processors.
We currently have 2 High-End GC's dedicated to Exchange 2003 - 8 Procs so we fall within the range of the 4:1 recommended ratio.
There are E5.5 boxes in the primary site but I have not added them to the equation as of yet.
Many of the settings are default with the exception of the relay settings which are restricted by subnet and the size restrictions which we use a standard of 5120kb.
The Issue -
During peak hours the awaiting directory lookup queues are starting to fill up. All perf counters return solid values, no event events are being generated that really point to a specific issue ... I've tried adding more GC's (2-3) and still had the same results.
I didn't have this issue until I upgraded to SP1 - I also rebuilt one of the IMS servers the same night. Now I'm having issues with all of the servers.
I apologize if this isn't enough info ... I'm more or less looking for potential issues I should look into, rather than a solution (unless you have one!) :)
Thanks,
Brian
ASKER
DCDiag from GC's looks good - NetDiag from IMS servers looks good - Cat Queue Lengths at 0 - Messages primarily hanging up in Waiting for Directory lookup queue on third IMS server (Server C) with a cost of 10 ... Logging turned up to Max on Transport, MTA and DSAccess ... only one recorded error -
Event Type: Warning
Event Source: MSExchangeTransport
Event Category: Categorizer
Event ID: 6004
Date: 7/28/2004
Time: 12:45:53 PM
User: N/A
Computer: CDC01IMS03
Description:
The categorizer is unable to categorize messages due to a retryable error. This operation returned because the timeout period expired.
I'm still at a loss ... all connectivity seems to be fine ... ugg!
Event Type: Warning
Event Source: MSExchangeTransport
Event Category: Categorizer
Event ID: 6004
Date: 7/28/2004
Time: 12:45:53 PM
User: N/A
Computer: CDC01IMS03
Description:
The categorizer is unable to categorize messages due to a retryable error. This operation returned because the timeout period expired.
I'm still at a loss ... all connectivity seems to be fine ... ugg!
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
A/P on the cluster, yes?
D
D
The store on Server C...is it mounted? If not, it would generate the error you posted with an SMTP connector....
d
d
ASKER
All Stores are mounted - no mailboxes on the Internet Mail Connector servers (SMTP Connectors) - the error is generated on the IMS Server C - the servers are still transferring email but the messages are queuing up in the "Awaiting directory lookup" queue.
GC Performance seems to be very good - servers aren't even reaching 1/5 th max load ...
Not sure what you mean by A/P on the cluster ...
GC Performance seems to be very good - servers aren't even reaching 1/5 th max load ...
Not sure what you mean by A/P on the cluster ...
active/passive....
D
D
ASKER
sorry ... yes AP ...
ASKER
So it turns out the issue was caused by a new feature enabled in our AV Software - Sybari Antigen 8.0 enables "Reverse DNS Lookups" by default .. given we pass about 200k messages through our IMC's each day this was causing latency on the lookups ...
The solution was to disable "Reverse DNS" in General Options of the Antigen Software on all of my E2K3 IMC's ....
The solution was to disable "Reverse DNS" in General Options of the Antigen Software on all of my E2K3 IMC's ....
good GOD! What a terrible option to set as default! You should throw rocks at them for that one....it was DNS, but not the way I thought.
D
D
ASKER
I forgot to mention ... if I use "Force Connection" it will start flushing the messages ... mail will appear to route properly for 5-10 minutes and then messages start queuing on the IMS servers ...