Flashcards in Transport Services Deck (175)
What is Back Pressure?
A state of resource exhaustion for Exchange servers, which leads to the server actively refusing some or all connections or e-mail delivery attempts.
It is a symptom of server sizing or performance issues. A server goes into a Back Pressure state as a method of self-preservation.
What are common causes of Back Pressure?
• Low free disk space on the drives that store the Transport Queue database and logs
• Too many uncommitted database transactions in memory
• Excess memory utilization, either by the transport service processes, or overall by the server
What Back Pressure States are there, and what do they indicate?
• no problems
• means a resource is moderately over-utilized
• server begins limiting some connection types, e.g. external connections, but continues processing internal mail
• means a resource is severely over-utilized
• server stops accepting all new connections
What are the signs of Back Pressure?
• SMTP error 452 4.3.1 Insufficient system resources
• Application event log entries, including ID #s 15004 through 15007.
What should you do if you discover your domain is listed on an anti-spam block list?
• Check block lists on a site such as MXToolbox.com to see what lists you appear on.
• See if lists provide a reason for your entry. Make corrections as required.
• Check that your receive connectors are properly configured and are not being used as an open relay for spammers.
• Contact the list that you have been blocked on to request removal.
What are Mail Flow Rules also known as?
What are Transport Rules?
• Rules created to look for messages matching specified criteria or conditions, and to take specified action on those messages.
• Applied to entire organization / at the organization level.
What are some examples of possible Transport Rule conditions?
• Sender details
– e.g. domain name
• Recipient details
– e.g. name, group membership, AD attributes)
• Message content
– e.g. keywords in subject, body, sensitive data types, attachments)
• Message size
– total size, attachment size, etc.
What are some examples of possible Transport Rule Actions?
• Forward or redirect message
• Reject or delete message
• Add additional recipients in To/Cc/Bcc
• Apply rights management/encryption
• Modify spam score
• Set any message header values
• Send message to a moderator for approval before sending out
• Apply a disclaimer to the message
What is a Shared SMTP Namespace?
An e-mail domain that is used by more than one mail organization / system, where each has a separate set of recipients using that domain.
What is required for setting up a Shared SMTP namespace?
• The domain added as an Accepted Domain (in each organization)
• Its Domain Type must be set to Internal Relay (in each organization)
• If a Recipient Filter is in place, the AddressBookEnabled property of the Domain must be set to $false, so that Recipient Filtering is not performed on it
• A send connector for that address space, set to route to the other organizations that share the namespace (in each organization)
• Transport Rules to prevent looped relaying of a message when a recipient cannot be found
How are loops prevented when a recipient cannot be found in a Shared SMTP Namespace?
• Exchange will automatically detect a loop and reject a message, but it takes about 30 loops before that detection kicks in.
• This can create delays and perhaps excessive traffic, so it's best to configure transport rules to reject a message after a single loop:
• Two Transport Rules are required:
– One rule tags the message header when a message is relayed out
– A second rule (with lower priority and thus and processed first) looks for the tag when receiving a message, and rejects it to prevent a loop
What is the default Message Size Limit?
For total message size, the default limit is 10 MB.
What is the default Message Recipient Limit?
For number of recipients, the default limit is 500.
How is a Distribution Group counted against Message Recipient Limits?
It is counted as a single recipient, even if the group has thousands of members.
At what level(s) can Message Size Limits be set?
• For the Organization
• For a Connector (Send, Receive, and AD Site Connectors)
• For a server
• For a recipient (Mailboxes and Distribution Groups)
If there are different message size limits set at different levels, what limit is used?
The most restrictive limit wins.
If a Message Size Limit is set at the Server Level, what will that impact?
Only mobile and web users.
I.e., only messages going through that server via ActiveSync, Outlook on the Web, and Exchange Web Services.
How is it recommended to utilize mutliple send connectors?
• A single Send Connector per internet-facing site is appropriate, and you generally should not configure more.
• Send connectors should usually be configured with equal cost, so that each site's Public IP is kept in use so they gain reputation on the internet as a good sender for your domain.
What is a Send Connector Cost, and what impact does it have?
• Send Connectors are configured with a "cost" which is part of the least-cost route calculation.
• One site may actually send out through another site's send connector, if the other send connector and the AD Site Link leading to it total to a lower cost than it's own send connector.
• But, using unequal costs so that only one Public IP gets utilized for outgoing mail is not recommended.
How do you configure multiple inbound routes for the purpose of redundancy?
Create multiple MX records, pointing to different Public IP addresses of different servers/sites.
They can be configured either with:
• equal priority, which results in a generally random distribution between the two routes,
• or differing priorities, which results in the lower priority record generally being attempted first.
If a message becomes queued because its route is unavailable, when will routing be retried?
• Queues will automatically retry every 10 mintues.
• You can force a retry by running this cmdlet: Retry-Queue
What does this cmdlet do, and when is it useful?
• When routing topology changes (such as changing route cost or creating a new connector), new routes are only evaluated for new email messages.
• So if messages are already in the transport queue because they failed to be routed, they will be retried using the same, previously determined route.
• Retry-Queue forces a retry, and the -Resubmit forces re-evaluation of the queued messages for new routes.
When a server needs to be taken down for maintenance, what happens to messages in its transport queue, and what should be done?
• The messages will remain in that server's queue and will not reattempt delivery until the server is online again.
• So, prior to taking the server down, you should:
– Drain transport queues.
– Redirect remaining queued messages to another server.
– Note, Shadow Redundancy or Safety Net messages on the server are not redirected, and thus will still be at risk.