Exchange Server Sluggish

Recently (like today) I noticed that many of the employees Outlook programs were running pretty slow and even got hung up when searching through old emails and even just at random.

Lots of "trying to connect to server" messages.

Our network is doing fine and everything else seems to be running smoothly. I haven't noticed the issues much because my Outlook is set to cached exchange mode. Does that help with server performance/load?

What other things are good to check on or do to increase performance of exchange and hopefully reduce the load that the server has to take on.

I am applying updates and the like, but I doubt that will have much benefit. It is not a dedicated exchange server which I know is not optimal, but we are a small company so I can't just make this server a sole exchange server.

Any ideas would be great...I am trying to find some good ideas or procedures to implement in the future that will help lighten the system load.

what version of exhange are you using? you can always consider running a defrag against the exchange server using eseutil....BUT please be warned that a) this will take time and b) you could seriously mess your information store by playing with this util.

Have you put the /3GB switch into the boot.ini?

Forgot some details. This is a Windows 2003 Exchange server. All client machines are Win XP SP2. We have a total of about 60 client machines.

The only other issue I have heard of today with the exchange server is one outside worker who says he can't use his outlook to send email, just receive it. Could that be related at all?

I have tried using a exchange analyzer tool to show me what areas might need some fixes, but nothing really jumps off the report.

A few issues with permissions on Public Folder and the SSL certificate expired, but that was about it. Of course I am sure the two guys who have been here the longest and have mailboxes of 7-8GB is probably heavily weighing down on the server...however, it didn't seem to have much of a performance affect in the past...perhaps i should just twist there arms more and get them to archive their stuff.

I have not tried the /3GB switch....I'll give it a go...

Any other things that can plague a exchange server with drops in performance?

you might want to add a filter gateway ( filter spam , viruses , etc ) before reaching the exchange server ...
When a virus is detected on exchange it will slow the performance of the server it self , i remember i red an article about excluding some running services from daily scan as well ( store.exe ? ) not sure ...

limiting attachment sizes can help too .

I noticed this too with Outlook using the cached mode - especially with users with vast amounts of emails - it's because it is synching the local copy with the Exchange Server constantly - if you remove the cached mode option then it will resolve the problem, however the user will no longer have an offline copy should they be using a laptop etc.

Hi Blunted One,

My advice is to forget the rocket science and get back to basics to start with.

You say that:

Quote:

---

Recently (like today) I noticed that many of the employees Outlook programs were running pretty slow

---

The first thing that would suggest to me is that your machine is getting FULL.

I generally work on the principle that you should have no more than 80% of total capacity utilised. This applies both to partitions and the total drivespace.

My experience of database driven applications on Windows boxes is that the performance deterioration tends to be quite marked (noticeable), and stepwise. This seems to match your field observations?

I have also noticed that individual's performance experiences can be related to the amount of their allocated server space that is utilised, although that could just be a volume thing.

The basic problem is that without sufficient working space, efficiency tools such as defragmentation and database compaction don't work properly :(

Go check on how much space you have on your server ;)

As for the external worker................ no, I think that is a problem local to him, or they would all be complaining?

Is this a MS Small Business Server??

Be careful with the 3gig switch if it is.

The SBS is a fussy beast...have you turned up the exchange logging??

What are your DNS settings on the server...workstations?

Depending on your version of SBS and service pack level...you have a 16 gig limit on the exchange....and once its full it stops working.

also I would follow Nihils advice and check your disk space.....min 20% free...else MS servers and apps start to slow?

What 3rd party applications do you have installed on the server, such as AV...and what is the exchange analyser tool??

MLF

Quote:

---

Originally Posted by WolfeTone

I noticed this too with Outlook using the cached mode - especially with users with vast amounts of emails - it's because it is synching the local copy with the Exchange Server constantly - if you remove the cached mode option then it will resolve the problem, however the user will no longer have an offline copy should they be using a laptop etc.

---

You'd think that cached mode would make thing run more smoothly? I say this because the cilent is using the local copy *after* the messages have been downloaded. When they go to open an email, the email is already on the machine and it doesn't have to wait for the server. Granted, it does have to wait for the server to update the local copy and it might be a tiny bit slower after everyone arrives to work for the day. However, this should reduce the load on the server overall... after the initial surge of updating the local cache. If bandwith is a problem, set it to download headers only first, then once there is enough bandwidth, or a user wants to view an item, it is downloaded. You can control the client's settings in group policy based on sites or groups. http://support.microsoft.com/kb/870926

What kind of hardware are we talking about? How much memory? How large are your mailbox stores? How large are mailboxes? Do you have quotas? Set to what? How often do you do offline maintenance of the machine? (be sure to get a full and verified backup before doing this) Are your OS, logs and databases on different physical harddrives/volumes? Have you run exchange best practices analyzer? What other uses does this server have? How many NICs? Are you load balancing the NICs? What type of switch? You used to network performance issues... are we sure that was taken care of? You said that this server has other roles... what types of roles? Have you run the performance monitor to check for disk, memory or cpu utilization problems? Do you have a benchmark to compare it to?

MLF: The exchange BPA and troubleshooting tool are great tools to help you head off potential problems. Or, once you encounter a problem. I run it about once a month or so, just to be sure everything is going smoothly. One of those CYA things... if I should ever have a problem and mgt comes to me later down the road, I can show them my log book with the reports printed and dated. :)

Excellent

Thanks Phish

Quote:

---

You must spread some Reputation around before giving it to phishphreek again.

---

MLF

Just to add to the OP questions....if Exchange is acting up...it usually will spew errors in the Event Viewer.

Any errors in there??

MLF

Is this exchange standard or enterprise? If it is standard you need to be watching your total database size. It is limited to 16 gb in standard. If you have two mailboxes that are 6-8gb a piece, you could be close.

But that will actually keep the store from mounting, so probably not the cause of what you are seeing, but be aware of it.

I'm really suprised that nobody has mentioned doing performance monitoring. That is where I would start. Run ExBPA in health mode and have it do a two hour health check. It will check disk IO, CPU, and memory utilization. If you have a hardware issue that will probably find it.

You can also just follow the general MS troubleshooting performance guide, it tells you exactly what perf counters to monitor and what is acceptable levels.

I also would not use the /3gb switch unless you totally understand how that will impact kernel memory and the ramifications of that switch. You need to have more than 1GB of memory for it to be of use. ExBPA should have picked this up if you need to use it.

Really?

I do love all these answers :D

errr what about KISS?

The FIRST thing you do is check the physical capacity of the system.

If it is too full it will run slow and crap out........... EOS?

I guess there is something in the saying "If the only tool you have is a hammer, you will see every problem as a nail"?????????????????

Only when you know that it is not a simple physical limitation problem, do you advance to more sophisticated possibilities. ;)

Nihil- In my opinion using perfmons or the ExBPA health check will point out any obvious performance or physical capacity issues. That is why I was saying I would not just run out making changes without first trying to understand what exactly is happening. Honestly, given the number of questions and complaints around exchange here recently I think the most prudent step would be some exchange administration training for the OP. Just my take though.

mohaughn: I did recommend to run exBPA, troubleshooting tool and performance monitor in my post. It was just burried in a lot of "brainstorm troubleshooting"... lol That's exactly how I think when I just wake up and I'm drinking my coffee. I either brainstorm troubleshoot or play suduko... then I can get ready for the day.

mohaughn,

Quote:

---

That is why I was saying I would not just run out making changes without first trying to understand what exactly is happening. Honestly, given the number of questions and complaints around exchange here recently I think the most prudent step would be some exchange administration training for the OP. Just my take though.

---

I couldn't agree more! Sure all the suggestions are fine, but I just have this gut feeling that we have a sort of "bigger picture" issue here?

I am sort of imagining a mature installation with an "if it ain't broke, don't fix it" mentality. Everything works fine until you start to hit capacity thresholds, then everything falls apart in most unpredictable ways.

There certainly doesn't seem to be any kind of preventative maintenance philosopy in place?

Also, we are told that this is not a dedicated server............. I wonder how much that clouds the issue?

:confused: