Note: This is the last of a set of articles about troubleshooting outbound messaging issues. If you need or want to read from the beginning, start here.
Now that you have some knowledge of where you can look and what information you can find, how do you use it to troubleshoot? Let's go back to my original roadmap:
First, know the path a message will take to get to its destination
Second, determine how far along that path it got
Third, figure out why it stopped (or if it stopped)
What path should a message take? By now you should know whether Exchange will send outbound messages to a Smarthost or directly to the recipient system.
How far did it get? Did it reach the Exchange server? Check Message Tracking. If the message doesn't appear, then the client (typically Outlook) never delivered it successfully. Check Outlook - is it still in the Outbox?
Let's say the message shows up in Message Tracking. Did the message leave the Exchange server? In other words, does it report it was transferred through SMTP? If it does, it means Exchange delivered the message to the next hop.
If not, what is the last thing reported by Exchange? Check the Queues. Remember your routing configuration. Does Exchange send all messages to a Smarthost, or does it use DNS?
It's at this point you may want to verify DNS lookups and test communications with Telnet.
If Exchange delivered the message to the next hop, it's time to examine the SMTP log. Checking the SMTP log will show you the communication between your Exchange server and the system to which it wanted to deliver the message.
Did it receive an OK response to the HELO/EHLO command?
Did it receive an OK response to the MAIL FROM command?
Did it receive an OK response to the RCPT TO command?
Did it receive an OK response to the DATA command?
Did it receive an OK response to the QUIT command?
If it passed all the way through the QUIT command, the message is now the responsibility of the system that received it from your Exchange server. If that system is under your administration, check there. If not, your troubleshooting has come to an end. You have verified that your system delivered the message successfully.
While this does not account for many situations, it does take a lot of the mystery out of troubleshooting. You can certainly dig deeper into SMTP and other (non-MS Exchange) systems, but I think this will start you on the way to becoming a troubleshooting guru. With an understanding of some of the concepts, you can read through technical articles and reference materials for more information.
Showing posts with label Exchange. Show all posts
Showing posts with label Exchange. Show all posts
Sunday, February 21, 2010
Understanding basic SMTP commands and responses
There are only a few SMTP commands commonly used, and more importantly only a few responses that matter.
SMTP responses
Recipient systems will respond to each SMTP command with a numeric value and optional text. Any value in the 200-299 range is considered to be an "OK" acknowledgement.
HELO / EHLO
This is how the sending system opens an SMTP conversation with a recipient system that acknowledges a TCP 25 communication attempt. HELO is the original SMTP specification, EHLO is an ESMTP command. Parameters after the command are optional, although it should be noted that some recipient systems may attempt to match that against the domain name indicated by performing a reverse-DNS lookup of the sending IP address.
MAIL FROM:
This command displays the reply address of the sender.
RCPT TO:
This command displays the recipient address. Only one address is allowed per command, so messages with multiple recipients will show each separately.
DATA
This command signifies the start of the actual message. That includes what appears in the TO, CC, BCC, and Subject lines of the message. It includes the message body and attachments. None of that information is displayed in the SMTP log. ESMTP sending systems may declare the length of the message.
QUIT
This command requests a termination to the SMTP session.
SMTP responses
Recipient systems will respond to each SMTP command with a numeric value and optional text. Any value in the 200-299 range is considered to be an "OK" acknowledgement.
HELO / EHLO
This is how the sending system opens an SMTP conversation with a recipient system that acknowledges a TCP 25 communication attempt. HELO is the original SMTP specification, EHLO is an ESMTP command. Parameters after the command are optional, although it should be noted that some recipient systems may attempt to match that against the domain name indicated by performing a reverse-DNS lookup of the sending IP address.
MAIL FROM:
This command displays the reply address of the sender.
RCPT TO:
This command displays the recipient address. Only one address is allowed per command, so messages with multiple recipients will show each separately.
DATA
This command signifies the start of the actual message. That includes what appears in the TO, CC, BCC, and Subject lines of the message. It includes the message body and attachments. None of that information is displayed in the SMTP log. ESMTP sending systems may declare the length of the message.
QUIT
This command requests a termination to the SMTP session.
Understanding the Queues view
For each server listed in the ESM, there is a folder called Queues. It displays the status of all connections the Exchange server is attempting to make. It will show the number of messages waiting to be sent. For any of the individual messages, you can see the sender, recipients, subject, and size. You can also see if a particular message is a Non-Delivery Report (NDR).
Messages that stay in the queue typically indicate a transmission problem. It could be that the destination domain doesn't exist, or that the destination domain is refusing your transmission, or that the destination domain is having a problem receiving messages. In any case, if a message stays in the queue, Exchange will attempt to send it again later.
Messages that stay in the queue typically indicate a transmission problem. It could be that the destination domain doesn't exist, or that the destination domain is refusing your transmission, or that the destination domain is having a problem receiving messages. In any case, if a message stays in the queue, Exchange will attempt to send it again later.
Understanding the SMTP log
Starting with Exchange 2003, all inter-server communication is by default done via SMTP. This makes the SMTP log a convenient way to see the high-level communication between your Exchange server and other SMTP systems. In particular, the SMTP commands and the responses to those commands.
Depending upon the amount of traffic your Exchange server handles, SMTP logs can get large. There is no automatic purging, so carefully consider where you are storing the log files. It is enabled/disabled on the General tab of the Default SMTP Virtual Server properties.
One of the unfortunate issues with the SMTP log is that there is no thread-organization. In other words, it is not possible to tell which log entry belongs to which thread. The entries are posted in the order received. If multiple threads are running concurrently, the entries will all be mixed together. That said, you can typically figure it out because of the sending address and recipient address.
Depending upon the amount of traffic your Exchange server handles, SMTP logs can get large. There is no automatic purging, so carefully consider where you are storing the log files. It is enabled/disabled on the General tab of the Default SMTP Virtual Server properties.
One of the unfortunate issues with the SMTP log is that there is no thread-organization. In other words, it is not possible to tell which log entry belongs to which thread. The entries are posted in the order received. If multiple threads are running concurrently, the entries will all be mixed together. That said, you can typically figure it out because of the sending address and recipient address.
Understanding Message Tracking
Message Tracking (MT) is a tool that is part the Exchange System Manager (ESM). It reports what a given Exchange server does with particular messages. Note that tracking ends when the message leaves that Exchange server. If the message goes to another Exchange server, you can consult MT on that other server for more information.
MT can be used to determine if a message was delivered to a mailbox, or to another system. If the tracking ends without the message being delivered, you will see what Exchange was doing last. That can give clues as to what the underlying issue is.
MT can be enabled or disabled. Open the properties of the Exchange server. On the General tab there is a checkbox to Enable Message Tracking. There are also settings for log location and retention.
MT can be used to determine if a message was delivered to a mailbox, or to another system. If the tracking ends without the message being delivered, you will see what Exchange was doing last. That can give clues as to what the underlying issue is.
MT can be enabled or disabled. Open the properties of the Exchange server. On the General tab there is a checkbox to Enable Message Tracking. There are also settings for log location and retention.
What's a Relay?
A Relay is essentially any system that accepts and forwards mail to SMTP domains for which it does not act authoritatively. That's a mouthful - what does that mean exactly? Let's say you address a message to uemurad@yahoo.com and send it to my corporate server. If my server accepts and forwards the message, my server is a relay. If my server accepts and forwards all messages from everyone, it's an open relay. Open relays are considered bad because Spammers find and utilize them to help disguise the source of the spam. Being an Open Relay is bad because recipient systems will see the spam coming from your system and will quickly degrade your reputation. You can end up on blacklists then have your legitimate messages rejected.
A system can be a Relay without being an Open Relay. A Relay is not a bad thing in itself.
A system can be a Relay without being an Open Relay. A Relay is not a bad thing in itself.
What's a Smarthost?
A Smarthost is essentially a system to which all messages are forwarded, regardless of their ultimate destination. Examples of Smarthosts are AntiVirus appliances, Content filters, Sender/Recipient filters, and Spam filters. In addition, some ISPs and some of their service levels require customers to send all outbound mail through their systems instead of directly to recipients. A Smarthost is responsible for receiving mail and forwarding to the appropriate system. Note that it too may send messages to a Smarthost or directly to a recipient system.
Understanding Logs
Logs are nothing more than a collection of status notes a system makes about a process. They are a way to review what happened during that process to understand what worked and what didn't. Most systems have some sort of logging option. Exchange has a couple of useful logging features that you've probably heard about - Message Tracking and SMTP logging.
Enabling Message Tracking is done in the Exchange System Manager (ESM). This is done at the server level. Open the properties of the server and find the checkbox on the General tab. That configuration also defines the location of the log files and the retention period.
Enabling SMTP logging is also done in the ESM. Although this is configured per server, the configuration is found in the properties of the Default SMTP Virtual Server (expand the server, Protocols, and SMTP).
Enabling Message Tracking is done in the Exchange System Manager (ESM). This is done at the server level. Open the properties of the server and find the checkbox on the General tab. That configuration also defines the location of the log files and the retention period.
Enabling SMTP logging is also done in the ESM. Although this is configured per server, the configuration is found in the properties of the Default SMTP Virtual Server (expand the server, Protocols, and SMTP).
Understanding your outbound mail flow
Before you can troubleshoot why outbound mail isn't getting to where it supposed to, you first have to understand how it's supposed to get there. One of the keys to this understanding is knowing whether or not Exchange sends messages directly to recipient systems. When an Exchange server gets a message destined for the outside world, there are two main mechanisms that it consults. The Virtual Server (VS) and Connectors.
The Virtual Server setting takes precedence, so look there first. Open the Exchange System Manager (ESM) and expand to the server in question. Further expand Protocols and SMTP. Beneath SMTP there is typically a single entry - the Default SMTP Virtual Server. Open the Properties of it, go to the Delivery tab, and click on the Advanced button. The main thing you are looking for is whether there is an entry in the Smart host field. If there is, it means that all outbound messages will be sent to that FQDN. If not, your Exchange Organization will require one or more SMTP Connectors to route outbound messages.
If the VS is not configured with a Smarthost, Exchange then looks to the Connectors for outbound routing information. The Connector structure can be simple or complex, depending upon the size of your Exchange Organization and how your enterprise wants mail to flow. For this example, let's assume the simplest configuration possible - a single Exchange server environment.
To route outbound messages, your single-server Exchange Organization requires at least one Connector (assuming you did not configure a smarthost via the VS).
Let's take a moment to review the function and difference between Exchange Administrative Groups (AG) and Exchange Routing Groups (RG). AGs allow you to easily configure administrative permissions to a group of servers, regardless of their geographic location or purpose. RGs are specifically for message routing. There is no implied correlation between any particular AG and RG.
Connectors are associated with RGs and are also referred to as SMTP Connectors.
Getting back our discussion on mail flow, in the simplest configuration you'd have a single Connector. Looking at the Properties, on the General tab there is a radio-button selection allowing you to either configure a smarthost or to use DNS.
If smarthost is selected, there should be an IP address in the field (as opposed to the FQDN format used in the VS).
If DNS is selected, it means that Exchange will attempt to send messages directly to the recipient system.
The Virtual Server setting takes precedence, so look there first. Open the Exchange System Manager (ESM) and expand to the server in question. Further expand Protocols and SMTP. Beneath SMTP there is typically a single entry - the Default SMTP Virtual Server. Open the Properties of it, go to the Delivery tab, and click on the Advanced button. The main thing you are looking for is whether there is an entry in the Smart host field. If there is, it means that all outbound messages will be sent to that FQDN. If not, your Exchange Organization will require one or more SMTP Connectors to route outbound messages.
If the VS is not configured with a Smarthost, Exchange then looks to the Connectors for outbound routing information. The Connector structure can be simple or complex, depending upon the size of your Exchange Organization and how your enterprise wants mail to flow. For this example, let's assume the simplest configuration possible - a single Exchange server environment.
To route outbound messages, your single-server Exchange Organization requires at least one Connector (assuming you did not configure a smarthost via the VS).
Let's take a moment to review the function and difference between Exchange Administrative Groups (AG) and Exchange Routing Groups (RG). AGs allow you to easily configure administrative permissions to a group of servers, regardless of their geographic location or purpose. RGs are specifically for message routing. There is no implied correlation between any particular AG and RG.
Connectors are associated with RGs and are also referred to as SMTP Connectors.
Getting back our discussion on mail flow, in the simplest configuration you'd have a single Connector. Looking at the Properties, on the General tab there is a radio-button selection allowing you to either configure a smarthost or to use DNS.
If smarthost is selected, there should be an IP address in the field (as opposed to the FQDN format used in the VS).
If DNS is selected, it means that Exchange will attempt to send messages directly to the recipient system.
Troubleshooting Outbound Messaging
A couple of years ago I wrote a series of articles on what to do when users say inbound messages aren't getting to them. I am frequently asked for help figuring out why outbound messages seemingly don't get to where they are going.
Troubleshooting outbound mail flow is relatively simple (at least to me, but then again I'm a self-professed Messaging-geek). Whenever I bring up the subject, people look at me like I'm talking about something more (dark) art than science. I think that's only true until you gain an understanding of the science. With that in mind, here's my attempt to explain some of that science and hopefully put you on the path to becoming a troubleshooting wiz.
The easiest way to think about messaging is that it is a point-to-point transmission system. It starts somewhere and wants to go somewhere else. It may require only one tranmission (hop) or several. Each transmission succeeds or fails. Sounds simple so far, right?
Let me put it another way - here's how I look at troubleshooting message flow:
First, you have to know the path a message will take to get to its destination
Second, you have to determine how far along that path it got
Third, you have to figure out why it stopped (or if it stopped)
Sounds simple enough, but as the saying goes the devil is in the details. Towards that end, here are some separate articles about narrowing the focus of your search. I'm not sure the best way to follow along. As you get used to some of these ideas you may have to go back and forth. You may want to jump to the last article in the list and see how far you can get. I'll try to link up the articles as much as possible.
Understanding your outbound mail flow
Understanding Logs
What's a Smarthost?
What's a Relay?
Understanding Message Tracking
Understanding the Queues view
Understanding the SMTP log
Understanding basic SMTP commands and responses
Using NSLookup to verify recipient system address
Using Telnet to verify SMTP connectivity
Let's Troubleshoot (putting it all together) !!
Troubleshooting outbound mail flow is relatively simple (at least to me, but then again I'm a self-professed Messaging-geek). Whenever I bring up the subject, people look at me like I'm talking about something more (dark) art than science. I think that's only true until you gain an understanding of the science. With that in mind, here's my attempt to explain some of that science and hopefully put you on the path to becoming a troubleshooting wiz.
The easiest way to think about messaging is that it is a point-to-point transmission system. It starts somewhere and wants to go somewhere else. It may require only one tranmission (hop) or several. Each transmission succeeds or fails. Sounds simple so far, right?
Let me put it another way - here's how I look at troubleshooting message flow:
First, you have to know the path a message will take to get to its destination
Second, you have to determine how far along that path it got
Third, you have to figure out why it stopped (or if it stopped)
Sounds simple enough, but as the saying goes the devil is in the details. Towards that end, here are some separate articles about narrowing the focus of your search. I'm not sure the best way to follow along. As you get used to some of these ideas you may have to go back and forth. You may want to jump to the last article in the list and see how far you can get. I'll try to link up the articles as much as possible.
Understanding your outbound mail flow
Understanding Logs
What's a Smarthost?
What's a Relay?
Understanding Message Tracking
Understanding the Queues view
Understanding the SMTP log
Understanding basic SMTP commands and responses
Using NSLookup to verify recipient system address
Using Telnet to verify SMTP connectivity
Let's Troubleshoot (putting it all together) !!
Tuesday, October 23, 2007
NDRs - Cause for concern?
Among the more common questions I get from my users concerns Non-Delivery Reports (NDRs) they receive for messages they didn't send. Some users are understandably upset because many of the subject lines would lead you to question their character.
In addition, I see lots of questions from messaging administrators concerned that their environment is an open relay because either they are receiving a lot of NDRs from outside systems, or their own queues are clogged with NDRs to other SMTP domains.
Recently I posted this response at MSExchange.org to an Exchange administrator searching for a reasonable explanation for the flood of inbound NDRs his users were receiving.
===============
Consider this scenario:
1. I am a "Secret Creator of Unwanted Messages" (a.k.a. SCUM)
2. There is an SMTP domain named company.com
3. The address no.one@company.com does not exist
4. Your SMTP domain is unsuspecting-user.com
5. You have a user with the address someone@unsuspecting-user.com
6. I send out a message addressed from someone@unsuspecting-user.com and to no.one@company.com
The message goes out from my secret lair in a nearby septic tank out to the Internet.
The message then gets delivered to the server at company.com. That server accepts the message, attempts to find a match for no.one@company.com and discovers there is no such address. Being the RFC-compliant system it is, it dutifully creates and sends out an NDR to the sender. The trouble is, it thinks the sender is someone@unsuspecting-user.com and so sends the NDR there.
===============
This isn't to say that we shouldn't watch for signs that our system has been compromised, but more on that later...
In addition, I see lots of questions from messaging administrators concerned that their environment is an open relay because either they are receiving a lot of NDRs from outside systems, or their own queues are clogged with NDRs to other SMTP domains.
Recently I posted this response at MSExchange.org to an Exchange administrator searching for a reasonable explanation for the flood of inbound NDRs his users were receiving.
===============
Consider this scenario:
1. I am a "Secret Creator of Unwanted Messages" (a.k.a. SCUM)
2. There is an SMTP domain named company.com
3. The address no.one@company.com does not exist
4. Your SMTP domain is unsuspecting-user.com
5. You have a user with the address someone@unsuspecting-user.com
6. I send out a message addressed from someone@unsuspecting-user.com and to no.one@company.com
The message goes out from my secret lair in a nearby septic tank out to the Internet.
The message then gets delivered to the server at company.com. That server accepts the message, attempts to find a match for no.one@company.com and discovers there is no such address. Being the RFC-compliant system it is, it dutifully creates and sends out an NDR to the sender. The trouble is, it thinks the sender is someone@unsuspecting-user.com and so sends the NDR there.
===============
This isn't to say that we shouldn't watch for signs that our system has been compromised, but more on that later...
Thursday, August 30, 2007
GAL vs. OAB - Why can't I see my AD changes?
When you run Outlook in non-cache mode, or connect using Outlook Web Access (OWA), you access the Global Address List (GAL) directly. Any changes made will appear as soon as your AD forest is replicated.
When you run Outlook in cache mode, you are viewing an offline copy of the GAL called an offline address book (a.k.a. offline address list). By default, Exchange rebuilds the OAB once a day. Also by default, Outlook downloads the OAB once when you launch it - thinking it doesn't need to check more often because of the rebuild schedule.
If you need to force a new entry to show up immediately, you need to do two things. First, manually rebuild the OAB. Open the System Manager (ESM) and expand Recipients then click on Offline Address Lists. In the right pane right click Default Offline Address List and select Rebuild. You should then wait anywhere from 2 to 10 minutes (depending upon how many entries are in your GAL). Second, in Outlook (running in cache mode) go to Tools-->Send/Receive and select Download Address Book. You should then see the new entry.
If you wish to view or change the rebuild schedule, open the ESM and expand Recipients. Open the properties of the particular OAB in question. On the General tab, there is a field for Update interval. By default this is set to run daily - early in the morning. If you wish to have the OAB rebuild more than once a day, you can select "Use custom schedule" and create your schedule.
When you run Outlook in cache mode, you are viewing an offline copy of the GAL called an offline address book (a.k.a. offline address list). By default, Exchange rebuilds the OAB once a day. Also by default, Outlook downloads the OAB once when you launch it - thinking it doesn't need to check more often because of the rebuild schedule.
If you need to force a new entry to show up immediately, you need to do two things. First, manually rebuild the OAB. Open the System Manager (ESM) and expand Recipients then click on Offline Address Lists. In the right pane right click Default Offline Address List and select Rebuild. You should then wait anywhere from 2 to 10 minutes (depending upon how many entries are in your GAL). Second, in Outlook (running in cache mode) go to Tools-->Send/Receive and select Download Address Book. You should then see the new entry.
If you wish to view or change the rebuild schedule, open the ESM and expand Recipients. Open the properties of the particular OAB in question. On the General tab, there is a field for Update interval. By default this is set to run daily - early in the morning. If you wish to have the OAB rebuild more than once a day, you can select "Use custom schedule" and create your schedule.
Saturday, July 14, 2007
Lesson learned - Recovering from corruption
As if the previous lesson learned wasn't enough, it led to another painful lesson.
Our Microsoft PSS engineer had us take a backup of the current mailbox stores that wouldn't mount along with the transactional logs so that we'd have something to fall back on should things not go well. Our mailbox stores are large enough that the process would take several hours. Rather than have new mail bounce he had us create empty databases by configuring the file names different from the original files. This gave the new messages a place to go.
After the backup completed, we restored one mailbox store files back to the previous day (which took a couple of hours). We then went through several rounds of replaying the logs until we figured out where the corruption started. When that process finally completed, we had a restored mailbox store as good as we could get it, plus a newly created mailbox store with a couple of days worth of messages. The final step is to merge the two. This is done by making one of the stores the Recovery Storage Group and merging the new data with the old.
I'll interrupt the story here to say that because of the length of time involved with the process, we ended up doing this with two different groups of people. The first group completed the merge and all appeared well. The second group went through the same process with another mailbox store with a different PSS engineer.
The bad news
When the second group got to the merge, it was taking a very long time. Much longer than it took the first group. We could do nothing but watch the merge wizard slowly process each mailbox. Our mailbox store was large enough that the entire process took over a day to complete. Our users were patient and understanding, but at the same time displeased.
Lesson learned
Some of you may have recognized that what we were doing was performing a Dial-tone Restore. Henrik Walther wrote an excellent set of articles on this subject at MSExchange.org. The first group did exactly what Henrik described, and the second group had left out one important step which would have saved us many hours of frustration. Before performing the merge, you need to swap the mailbox stores so that you are merging the small into the large. The mailbox store was somewhere between 75GB and 100GB and therefore took a very long time!
I strongly urge everyone to read through Henrik's articles to familiarize yourself with the process. You never know when you'll need it!
Our Microsoft PSS engineer had us take a backup of the current mailbox stores that wouldn't mount along with the transactional logs so that we'd have something to fall back on should things not go well. Our mailbox stores are large enough that the process would take several hours. Rather than have new mail bounce he had us create empty databases by configuring the file names different from the original files. This gave the new messages a place to go.
After the backup completed, we restored one mailbox store files back to the previous day (which took a couple of hours). We then went through several rounds of replaying the logs until we figured out where the corruption started. When that process finally completed, we had a restored mailbox store as good as we could get it, plus a newly created mailbox store with a couple of days worth of messages. The final step is to merge the two. This is done by making one of the stores the Recovery Storage Group and merging the new data with the old.
I'll interrupt the story here to say that because of the length of time involved with the process, we ended up doing this with two different groups of people. The first group completed the merge and all appeared well. The second group went through the same process with another mailbox store with a different PSS engineer.
The bad news
When the second group got to the merge, it was taking a very long time. Much longer than it took the first group. We could do nothing but watch the merge wizard slowly process each mailbox. Our mailbox store was large enough that the entire process took over a day to complete. Our users were patient and understanding, but at the same time displeased.
Lesson learned
Some of you may have recognized that what we were doing was performing a Dial-tone Restore. Henrik Walther wrote an excellent set of articles on this subject at MSExchange.org. The first group did exactly what Henrik described, and the second group had left out one important step which would have saved us many hours of frustration. Before performing the merge, you need to swap the mailbox stores so that you are merging the small into the large. The mailbox store was somewhere between 75GB and 100GB and therefore took a very long time!
I strongly urge everyone to read through Henrik's articles to familiarize yourself with the process. You never know when you'll need it!
Friday, July 13, 2007
Addendum - Stopping Exchange services
Andy Grogan wrote an excellent commentary to my previous post, and I wrote in return. I don't know how many people would actually read the comments, and the points made are important enough that I wanted them to be more than a footnote. I left Andy's comments verbatim, and added a couple of notes to mine.
Andy wrote:
"The problem is, how do you know when the store is being written to? or not, I have seen a similar problem before CPU 100 % nothing much responding in terms of Exchange so you naturally try to stop the service. Then you get the dreaded "The Microsoft Information Store Service did not respond to the stop request in a timely fashion" and hang on stopping. This then raises - how long do you wait - and hour, two hours - a day?"
"I have in the past opened up Task Manager and added in the IO, and Bytes counters to see if the store process is writing, but its not fool proof. Personally I would love Microsoft to put some options into ending processes like there are in Unix where you have multiple levels to killing a process - sorry for the ramble - I just don't think there was much you could have done your were in a hole that most of us Exchange folks face at some point - do I pull the plug or don't I?"
The issues Andy raises in his comment are exactly the same questions and issues that haunted me. That's why I've adopted this new strategy which seems to work (at least it alleviates my fears). Dismount the stores first - don't try to do anything with the services. In fact, don't do anything with the services until the stores dismount.
In an emergency, I'd pull the plug on new messages (e.g. block port 25) and cause bounces rather than risk corruption. That will allow the stores to eventually catch up and quiet down. Then dismount the stores, then finally stop the services.
Andy wrote:
"The problem is, how do you know when the store is being written to? or not, I have seen a similar problem before CPU 100 % nothing much responding in terms of Exchange so you naturally try to stop the service. Then you get the dreaded "The Microsoft Information Store Service did not respond to the stop request in a timely fashion" and hang on stopping. This then raises - how long do you wait - and hour, two hours - a day?"
"I have in the past opened up Task Manager and added in the IO, and Bytes counters to see if the store process is writing, but its not fool proof. Personally I would love Microsoft to put some options into ending processes like there are in Unix where you have multiple levels to killing a process - sorry for the ramble - I just don't think there was much you could have done your were in a hole that most of us Exchange folks face at some point - do I pull the plug or don't I?"
The issues Andy raises in his comment are exactly the same questions and issues that haunted me. That's why I've adopted this new strategy which seems to work (at least it alleviates my fears). Dismount the stores first - don't try to do anything with the services. In fact, don't do anything with the services until the stores dismount.
In an emergency, I'd pull the plug on new messages (e.g. block port 25) and cause bounces rather than risk corruption. That will allow the stores to eventually catch up and quiet down. Then dismount the stores, then finally stop the services.
Wednesday, July 11, 2007
Lesson Learned - Stopping Exchange services
I inadvertently corrupted a couple of mailbox stores (like anyone actually causes corruption intentionally) because of a lack of understanding. I hope this story will save someone else some pain.
The story started when an Exchange server's CPU was running at 100% for a long time, causing my monitor to alert us. After some intial troubleshooting, it was decided to restart the Exchange services. The restart process failed, the CPU continued to run at 100%.
We waited an hour, and the status had not changed. Figuring a reboot would resolve any ailments, we did just that. The server took a long time, but finally rebooted after 15 minutes. When the server restarted, some of the mailbox stores did not mount. After spending some time to try to fix things, we put in a call to Microsoft PSS.
The bad news
We discovered that our actions corrupted several mailbox stores (this was on an Enterprise Edition server). In talking to Microsoft, we discovered that a shutdown or restart of the Operating System does not necessarily wait for all services to stop. The Information Store service apparently did not stop completely and after a timed delay, Windows shut itself down. We were told that the corruption happened because the store was actively being written to when the service stopped.
Lesson learned
With a new understanding of the Information Store service, whenever maintenance is performed on our Exchange servers, we always dismount all of the mailbox stores first. This assures that all "in flight" transactions are complete before the service is stopped.
The story started when an Exchange server's CPU was running at 100% for a long time, causing my monitor to alert us. After some intial troubleshooting, it was decided to restart the Exchange services. The restart process failed, the CPU continued to run at 100%.
We waited an hour, and the status had not changed. Figuring a reboot would resolve any ailments, we did just that. The server took a long time, but finally rebooted after 15 minutes. When the server restarted, some of the mailbox stores did not mount. After spending some time to try to fix things, we put in a call to Microsoft PSS.
The bad news
We discovered that our actions corrupted several mailbox stores (this was on an Enterprise Edition server). In talking to Microsoft, we discovered that a shutdown or restart of the Operating System does not necessarily wait for all services to stop. The Information Store service apparently did not stop completely and after a timed delay, Windows shut itself down. We were told that the corruption happened because the store was actively being written to when the service stopped.
Lesson learned
With a new understanding of the Information Store service, whenever maintenance is performed on our Exchange servers, we always dismount all of the mailbox stores first. This assures that all "in flight" transactions are complete before the service is stopped.
Tuesday, June 26, 2007
Relays - good or bad?
I see many questions about relaying. It seems much of the confusion stems from an incomplete or incorrect understanding of what it means. Hopefully this will clear things up a bit.
What is relaying?
For any given message, you have a sending system (originator) and a recipient system. Most of the time, the sending system communicates directly with the recipient system. So, if you were sending a message to me, your Exchange server would communicate directly with mine.
There are times and circumstances where direct communication is not possible or desired. These circumstances may require that an intermediate system field the message and pass it along. This is what a relay is. It is a neither the sending system nor the recipient system.
Many ISPs work this way where you have to send all your outbound mail to them to be sent out to other Internet domains. The ISP accepts messages only from its known customers (via IP address) and rejects messages from others. Another common example of a valid relay is a Spam Filter appliance.
Open relays
A big problem that we Messaging Admins face is that there are "open relays" out in the world. An Open Relay is when a system accepts messages from everyone and will forward them to anyone. Spammers love these system because they can hide behind open relays to mask the true originating system. The default configuration of Exchange 5.5 unfortunately was set up to be an open relay. Many mail systems were unknowningly left that way, creating a paradise for spammers. Relaying with Exchange 2000 by default was closed to all but authenticated users, and Exchange 2003 continues that same default configuration.
Relay settings in Exchange 2003
These are set in the properties of the Default SMTP Virtual Server, on the Access tab.
If you select "all except the list below" in the relay settings with a blank list, you are actually saying "forward messages from everywhere". In other words, you will have configured an open relay and raised the frustration level of legitimate Messaging Admins everywhere.
When you select "only the list below" with a non-blank list, you are saying "don't forward the message unless it comes from a system on the list.
A final important configuration note
If you clear the checkbox for "Allow all computers which successfully authenticate to relay regardless of the list above", then Exchange servers within your organization will not be able to send messages to one another.
What is relaying?
For any given message, you have a sending system (originator) and a recipient system. Most of the time, the sending system communicates directly with the recipient system. So, if you were sending a message to me, your Exchange server would communicate directly with mine.
There are times and circumstances where direct communication is not possible or desired. These circumstances may require that an intermediate system field the message and pass it along. This is what a relay is. It is a neither the sending system nor the recipient system.
Many ISPs work this way where you have to send all your outbound mail to them to be sent out to other Internet domains. The ISP accepts messages only from its known customers (via IP address) and rejects messages from others. Another common example of a valid relay is a Spam Filter appliance.
Open relays
A big problem that we Messaging Admins face is that there are "open relays" out in the world. An Open Relay is when a system accepts messages from everyone and will forward them to anyone. Spammers love these system because they can hide behind open relays to mask the true originating system. The default configuration of Exchange 5.5 unfortunately was set up to be an open relay. Many mail systems were unknowningly left that way, creating a paradise for spammers. Relaying with Exchange 2000 by default was closed to all but authenticated users, and Exchange 2003 continues that same default configuration.
Relay settings in Exchange 2003
These are set in the properties of the Default SMTP Virtual Server, on the Access tab.
If you select "all except the list below" in the relay settings with a blank list, you are actually saying "forward messages from everywhere". In other words, you will have configured an open relay and raised the frustration level of legitimate Messaging Admins everywhere.
When you select "only the list below" with a non-blank list, you are saying "don't forward the message unless it comes from a system on the list.
A final important configuration note
If you clear the checkbox for "Allow all computers which successfully authenticate to relay regardless of the list above", then Exchange servers within your organization will not be able to send messages to one another.
Saturday, June 9, 2007
Using Telnet to simulate server communication
The best place to run Telnet is on the server which sends out your SMTP traffic. This will show you the same information that your SMTP engine receives when communicating with an outside system. Telnet allows you to specify the port through which to communicate. SMTP is defined as TCP port 25.
Open a command prompt window. Determine the FQDN or the IP address. If you need to determine this information, you can use NSLookup if you know the SMTP domain name you are attempting to connect to. For more information about this, read this article.
At the prompt, type telnet fqdn 25
If the receiving server is accepting SMTP communications, it will respond with an acknowledgement message indicating it is ready to receive your transmission. The acknowledgement should also indicate if it understands SMTP or ESMTP.
Type ehlo testdomain.com
There are two established protocols, SMTP and ESMTP (Enhanced SMTP). If the receiving system only understands SMTP, you must begin with helo. If the receiving system understands ESMTP, you may begin with either helo or ehlo.
If you receive an OK message from the receiving mail system, proceed. If not, double check the protocol named in the response to the telnet command.
Type
mail from:exchange.admin@testdomain.com
This indicates the reply address. Some receiving systems will compare the parameter from the ehlo command, and the domain listed in the address on the mail from: command to the domain name returned when performing a reverse DNS (RDNS) lookup on the IP address from which the message is coming. It is a method to combat address spoofing and more reliably identify undesirable senders.
If you are testing communications to an outside messaging system, you may need to use your actual domain name to be allowed to continue.
Type
rcpt to:valid.user@receivingdomain.com
This indicates the recipient address. If receivingdomain.com is not a domain being fielded by the receiving system, and the system does not allow relaying to receivingdomain.com, an error code will be returned.
Type data
This begins the actual message. Optionally, From:, To:, and BCC: can be entered at this time (to be covered in a future article).
Type subject: Test message via Telnet
Type a blank line - this denotes the end of the subject and the beginning of the message body.
Type This is a test
Type Please reply if received
Type a blank line
Type a period (".") and press Enter - this marks the end of the message body. The receiving system will understand and return a prompt.
Type quit
This ends the Telnet session and you will be returned to the OS prompt.
If everything has gone well, the message will be on its way to the recipient address. Give it a minute and check. You now know how to manually create and send an SMTP message! This can be a great troubleshooting tool, as you will receive reponses and acknowledgements from the receiving system that can aid in diagnosing a communication problem.
Open a command prompt window. Determine the FQDN or the IP address. If you need to determine this information, you can use NSLookup if you know the SMTP domain name you are attempting to connect to. For more information about this, read this article.
At the prompt, type telnet fqdn 25
If the receiving server is accepting SMTP communications, it will respond with an acknowledgement message indicating it is ready to receive your transmission. The acknowledgement should also indicate if it understands SMTP or ESMTP.
Type ehlo testdomain.com
There are two established protocols, SMTP and ESMTP (Enhanced SMTP). If the receiving system only understands SMTP, you must begin with helo. If the receiving system understands ESMTP, you may begin with either helo or ehlo.
If you receive an OK message from the receiving mail system, proceed. If not, double check the protocol named in the response to the telnet command.
Type
mail from:exchange.admin@testdomain.com
This indicates the reply address. Some receiving systems will compare the parameter from the ehlo command, and the domain listed in the address on the mail from: command to the domain name returned when performing a reverse DNS (RDNS) lookup on the IP address from which the message is coming. It is a method to combat address spoofing and more reliably identify undesirable senders.
If you are testing communications to an outside messaging system, you may need to use your actual domain name to be allowed to continue.
Type
rcpt to:valid.user@receivingdomain.com
This indicates the recipient address. If receivingdomain.com is not a domain being fielded by the receiving system, and the system does not allow relaying to receivingdomain.com, an error code will be returned.
Type data
This begins the actual message. Optionally, From:, To:, and BCC: can be entered at this time (to be covered in a future article).
Type subject: Test message via Telnet
Type a blank line - this denotes the end of the subject and the beginning of the message body.
Type This is a test
Type Please reply if received
Type a blank line
Type a period (".") and press Enter - this marks the end of the message body. The receiving system will understand and return a prompt.
Type quit
This ends the Telnet session and you will be returned to the OS prompt.
If everything has gone well, the message will be on its way to the recipient address. Give it a minute and check. You now know how to manually create and send an SMTP message! This can be a great troubleshooting tool, as you will receive reponses and acknowledgements from the receiving system that can aid in diagnosing a communication problem.
Friday, June 8, 2007
Using NSLookup to determine an SMTP receiving system
Background
NSLookup is a great tool that comes with Windows that allows you to search DNS for information.  It is especially useful to troubleshoot particular issues with Exchange.  Exchange is reliant upon DNS to know where to send outbound messages.  When Exchange has problems getting messages to a particular domain, it's time to open the toolbox.
The best place to run NSLookup is on the server which sends out your SMTP traffic.  This will show you the same information that your SMTP engine uses when determining where to send mail to a particular domain.
Open a command prompt window
At the prompt, type nslookup
Type the command set type=mx
Type the registered domain name (e.g. domain.com)
You will receive a response similar to:
Non-authoritative answer:
domain.com MX preference = 10, mail exchanger = mail1.domain.com
domain.com MX preference = 20, mail exchanger = mail2.domain.com
domain.com MX preference = 30, mail exchanger = mail3.domain.com
Interpreting the NSLookup results
Your SMTP engine will attempt to use the MX records in ascending order according to their value.  The name associated with the MX record is what your engine will use.  You can simulate what the engine does by using the Telnet command.  In other words, the FQDN associated with the lowest numbered MX value would be the one that your SMTP engine would attempt to connect with.
Using the NSLookup results to test connectivity
In the simulated response shown above, you can test the readiness for receiving SMTP communications by using the Telnet command.  In a command-prompt window, type telnet mail1.domain.com 25.  If the system connected to the FQDN is accepting SMTP communications, you’ll receive a response.
NSLookup is a great tool that comes with Windows that allows you to search DNS for information.  It is especially useful to troubleshoot particular issues with Exchange.  Exchange is reliant upon DNS to know where to send outbound messages.  When Exchange has problems getting messages to a particular domain, it's time to open the toolbox.
The best place to run NSLookup is on the server which sends out your SMTP traffic.  This will show you the same information that your SMTP engine uses when determining where to send mail to a particular domain.
Open a command prompt window
At the prompt, type nslookup
Type the command set type=mx
Type the registered domain name (e.g. domain.com)
You will receive a response similar to:
Non-authoritative answer:
domain.com MX preference = 10, mail exchanger = mail1.domain.com
domain.com MX preference = 20, mail exchanger = mail2.domain.com
domain.com MX preference = 30, mail exchanger = mail3.domain.com
Interpreting the NSLookup results
Your SMTP engine will attempt to use the MX records in ascending order according to their value.  The name associated with the MX record is what your engine will use.  You can simulate what the engine does by using the Telnet command.  In other words, the FQDN associated with the lowest numbered MX value would be the one that your SMTP engine would attempt to connect with.
Using the NSLookup results to test connectivity
In the simulated response shown above, you can test the readiness for receiving SMTP communications by using the Telnet command.  In a command-prompt window, type telnet mail1.domain.com 25.  If the system connected to the FQDN is accepting SMTP communications, you’ll receive a response.
Thursday, June 7, 2007
DNS Root Servers
What are the root servers? These are the DNS servers from which all others get information.
Why would anyone care? To verify the information other mail systems see when trying to reach your SMTP domain.
How do I use the information? You can force NSLookup to poll a particular DNS server by using the command:
server ip_address or server fqdn
As an example, open a command prompt window
At the prompt, type nslookup
Type the command set type = mx
Type the command server a.root-servers.net
Type the registered domain name (e.g. domain.com)
You have requested the MX information for domain.com directly from one of the DNS root servers.
Here is IP information for the thirteen (13) root servers.
a.root-servers.net 198.41.0.4
b.root-servers.net 192.228.79.201
c.root-servers.net 192.33.4.12
d.root-servers.net 128.8.10.90
e.root-servers.net 192.203.230.10
f.root-servers.net 192.5.5.241
g.root-servers.net 192.112.36.4
h.root-servers.net 128.63.2.53
i.root-servers.net 192.36.148.17
j.root-servers.net 192.58.128.30
k.root-servers.net 193.0.14.129
l.root-servers.net 198.32.64.12
m.root-servers.net 202.12.27.33
The complete information can be found at the website: http://www.root-servers.org
Why would anyone care? To verify the information other mail systems see when trying to reach your SMTP domain.
How do I use the information? You can force NSLookup to poll a particular DNS server by using the command:
server ip_address or server fqdn
As an example, open a command prompt window
At the prompt, type nslookup
Type the command set type = mx
Type the command server a.root-servers.net
Type the registered domain name (e.g. domain.com)
You have requested the MX information for domain.com directly from one of the DNS root servers.
Here is IP information for the thirteen (13) root servers.
a.root-servers.net 198.41.0.4
b.root-servers.net 192.228.79.201
c.root-servers.net 192.33.4.12
d.root-servers.net 128.8.10.90
e.root-servers.net 192.203.230.10
f.root-servers.net 192.5.5.241
g.root-servers.net 192.112.36.4
h.root-servers.net 128.63.2.53
i.root-servers.net 192.36.148.17
j.root-servers.net 192.58.128.30
k.root-servers.net 193.0.14.129
l.root-servers.net 198.32.64.12
m.root-servers.net 202.12.27.33
The complete information can be found at the website: http://www.root-servers.org
Sunday, June 3, 2007
Missing messages - Part 4
The message was delivered to the mailbox - where did it go?
This is the most common scenario [I mean, speaking as one Exchange Admin to another - what else could it be? ;) ]. As a personal aside, your goal is to figure out what happened and calmly point it out to the user. The user will likely feel embarrassed already - no need to editorialize or lecture.
As discussed in Part 1, a successful message delivery typically means one of the following:
- It reached the mailbox and was segregated or deleted by a system function
- It reached the mailbox and was segregated or deleted by a client function
- It reached the mailbox and was segregated or deleted by a user function
- It reached the mailbox and was manually segregated or deleted
System function typically means forwarding configured in AD
To check for forwarding, open the Users & Computers console (ADUC) and open the properties of the recipient's object. On the Exchange General tab go to Delivery Options. Any forwarding configured at the Active Directory level will appear there.
Client functions include anti-virus/anti-spam filtering, and directing new messages to a personal folder
Check the console and logs of any 3rd-party anti-virus and anti-spam software. Check the Junk E-mail folder in the user's mailbox.
Check all workstations this user logs on to for a profile that directs all new messages to a Personal Folder instead of to the mailbox.
User functions include rules, auto-archiving, and viewing filters
Check for and disable any viewing filters in Outlook (View-->ArrangeBy-->Custom)
Check for auto-archiving (File-->Archive), look in all Personal folders listed in the Outlook profile. Search for all PST files on the local drive and all mapped drives.
Check for rules (Tools-->Rules and Alerts)
If the ruleset is empty, there is still a possibility that something formerly in rules is still acting on messages. To make sure, close Outlook, then launch it again from a command line using the /cleanrules switch (e.g. outlook.exe /cleanrules)
If the ruleset is not empty and you wish to keep them, you can export the set to a file then import again later.
Remember that the Out Of Office function can also have rules. If OOO is enabled, make sure you check that configuration for rules.
Manual processes initiated by the user
Look for and search any PST files in the Outlook profile and on the local drive.
Look in the Deleted Items folder. Look at the Recover Deleted Items area.
Search the other folders for items which were Shift-Deleted.
Subscribe to:
Posts (Atom)