Non-Delivery Messages 1

[proudusa]

I'm getting non-delivery for e-mails being sent out of outlook 2000 which are configured with an external POP/SMTP server in their properties. Mailboxes are stored on an Exchange 5.5 server. The server is used to store mail and do collaborative work but does not have IMS (SMPT).

Clients are able to receive e-mails but unable to send. They get: Your message did not reach some or all of the intended recipients.

Subject: test
Sent: 5/16/2003 3:22 PM

The following recipient(s) could not be reached:

'user@test.com' on 5/16/2003 3:22 PM
No transport provider was available for delivery to this recipient.

On the client side i'm seeing:

win98/ME: After exiting outlook following non-delivery receipt you get an error: MAPISP32.exe caused an invalid page fault in module rtfhtml.dll

Windows 2000: After exiting outlook following non-delivery receipt you get an error MAPISP32.EXE. The instruction at "....." referenced memory @ "...". The memory could not be read.

Have tried KB 318658 and KB 813741 with no luck. Problem must be central as it's all outlook clients that cannot send. Could be:

E-mail provider (says nothing going on there)
Firewall blocking (nothing has changed i'm not seeing anything specific logged)
Exchange server (only stores the mailboxes with mail and not used as SMTP or POP/IMAP server)

Please advice or point in right direction.

 

[mnkelznr]

My exchange 5.5 (sp4) running under sbs 4.5 (sp6) also began to have a very similar problem. Several of my exchange users have seperate "Internet E-mail" or pop accounts which they have delivered to thier exchange mailbox when outlook 2000 machines are attached to the lan and logged into the exchange server.

Several had been using the "Internet E-mail" or POP
(as in Tools, Services, Delivery )
accounts for delivery optiokn until 5/19, then bang it stopped working... I had to switch them to use "Exchange Transport" and again they are working.

I think this will cause trouble when they are not attached to LAN though, still wondering what happend. I suspect ISP has done something, as Exchange 5.5 has been running happily this way for a very long time.

Outlook stops responding or gets really slow when you try to send/rec if you end outlook imapisp32.exe on client computer seem to hang in task list, and store.exe on server seems to go nutz and eat up all the CPU on the server.

 

[mikewy]

OK-- Now I am becoming more suspicious. Three of us with similar problems within the same number of days

Mike

 

[Marcs41]

You would not have all 3 the same ISP would you?

 

[mikewy]

Wouldn't that be the grandest coincidence. Doubt it - I'm in Wyoming and isp is 180com.net. I gave up and am going to try to re-load the server tonight. Don't have any other server to move this stuff to at the moment. I'll post how it goes later.

Mike

 

[mnkelznr]

I'm using AMERITECH.net or sbcglobal for my access and webhosting services, I have tried to contact them, but they made several changes over the last year.

They changed their backbone from UUNET to sprint and a that and they must have fiddeled with their DNS servers, just recently because a few in my list dont work well anymore.

Who knows, the changes I made in outlook seem to have fixed the problems for now... I will worry about it on Tue I guess.

 

[twillboy]

That is really odd. I have the exact same setup w/ Exchange 5.5 on an NT box and using external pop3/smtp and the exchange server to house mailboxes. I get the same exact error message.

I'm in Iowa and I do know that our provider (Waverly Light and Power) do route through sprint because our DNS settings point to ns2.sprintlink.net and ns3.sprintlink.net

Our provider also just changed some stuff recently but it was on May 19th that we had users experiencing problems. The machine that I write this note from has Win95 w/ Outlook 98 and no mail can be sent outside the company here.

However the Win2000 machine in the corner from me w/ outlook 2000 has zero problems with anything. Nothing was changed on the outlook server and nothing was changed on the 30+ computers that are having problems.

Anyhow, again, I agree that it is really suspicious that all this happened on the 19th and Microsoft knows "jack" about it.

peace

 

[mikewy]

Oh my, another one with the same problem on the same day! Now I am convinced something devious is going on here.
Sounds like Proudusa found a fix by moving the mailboxes to another server. I'm trying to isolate what caused it. Here's what I've tried so far WITHOUT success:
-Uninstalling Exchange, re-installing it, and restoring the IS & DS
-Installing Exchange Service Pack 4 (only had SP 3 on it)

Tonight we are going to try stripping down the server completely, re-installing the OS & Exchange, then restoring the IS & DS. I'll post what happens.

Mike

 

[mikewy]

Progress (or not) report.
Stripped down the exchange server, wiped the drive and did a fresh install of Win 2000 server and Exchange. Then added some accounts and tested it with the internet e-mail. Everything worked as it should.
Then, restored the DS and IS data from backup, and the problem came back. The problem is obviously caused by something in either the Directory or Information store databases. Even deleted the mailboxes, one at a time to see if it would clear up the trouble with no luck. Out of ideas now, except to try Proudusa's fix.
This situation is really disturbing, and appears to be more wide-spread than we know. As I search the internet for possible fixes, I am finding more people with the same problem. And they occurred over May 18-19. What worries me more is, even if I get it fixed, I am thinking there is a good possibility that whatever caused it will happen again!

Mike

 

[ALAN1980]

My users started having this same problem on May 22 around 8:30am Pacific time. We can receive and send internal e-mail with no problem. We can even receive Internet e-mail without a problem. But when we try to send out Internet e-mail we get the "No transport..." error. At this point I think we might have reached the 16GB limit on the Exchange Server. Since I'm not a Exchange expert I do not how to confirm this. The total data files size shows 12GB. The current workaround for some of my users is to setup Outlook Express for all their Internet e-mail.

 

[jonarden]

The timing of the issue (May 16-22) impacting a very secific Exchange setup that ws working is interesting. Reviewing your posts it appears that:
a. It was not the result of any known change(s) anyone made to Exchange 5.5, server platforms or client machines.
b. Proudusa reports issue w/ users running Outlook 2000 on Win 2000, ME & 98; twilby running Outlook 98 on Win 95.
c. Occurs w/ different server platforms (mikewy running Exchange 5.5 on Win 2000 server, mnklzner on SBS, and twillboy on NT)
d. Occurs with different ISPs (mikewy 180com.net, mnkelzner Ameritech & sbcglobal w/ sprint backbone, twilboy Waverly Light and Power w/ sprint backbone.

The Fix proudusa implemented is restoring mailboxes to a new Exchange 5.5 install.

"Shots in the Dark" about cause, based on speculating about possible "automatic" or "background" events.
1. What anti-virus software is everyone using on Exchange 5.5? When were new definitions (or "super DATs) downloaded, and when were scans of full IS set to run during the time(s) this issue occurred? (check logs). Possible scan impact on the IS? Any virus (trojans) detected?
2. Check for trojans on impacted client machines. Run a full system scan on all impacted client machines and (perhaps) instll new OS and Outlook on a "test machine" configured for an exiting user milbox and see if POP3 now works. Make sure Exchange info store is also "clean" before you do this.
3. I assume automated MS "hotfixes" aren't an issue at your sites, since problem is occurring over broad range of server platforms and client OSs. BUT it just might be useful to push for specific answers from your ISPs about what they have done on their systems immediatly prior to the event (although this is a "dark horse", since proudusa has a fix that did not require any ISP action). Obviously I'm just speculating with all this, so apologies if I'm "out of line".

 

[ALAN1980]

- We are using Norton Corp. Anti-virus.
- All Outlook versions have been impacted. Outlook 98 and 2000.
- Event viewer shows that :

EventID: 183 Source ESE97
MSExchangeIS(293) Online defragmentation of database PRIV.EDB was interrupted and terminated

EventID: 1221 , Source: MSexchangeIS Public
The database has 2 megabytes of free space after online defragmentation has terminated

The above two see to happen everyday for the last week at 7:59:59am. Does anyone else have similar problems?

 

[ALAN1980]

I found the solution to:

EventID: 183 Source ESE97
MSExchangeIS(293) Online defragmentation of database PRIV.EDB was interrupted and terminated

It appears that you need at least 110% free space to defrag PRIV.EDB.

I also found out the sizes of my EDB files (I'm learning alot very fast):
PRIV.EDB 11.3GB PUB.EDB 491MB.

The drive in these two files are located on a drive that only has 15GB. Therefore Exchange cannot do the defrag process. One of my vendors contacted Microsoft Tech support for me and said the "No transport ..." error can be related to reaching the max size on PRIV & PUB files.

I'm going to try to fix this by ording a new 36GB drive. I'm also having all the users clean-out their boxes. Once I do this than I will install the new drive and move all the data to it. Once this is completed I'll try to defrag. Hopefully this will clear the "No transport..." my users are getting.

Sorry if I get to wordy.

 

[mikewy]

Well, no hard-drive shortage here. Defrag has plenty of room to work. Also, looking at the other posts, nothing was happening on our exchange server when it went nuts. Nothing in the event logs, and only 3 users logged onto it at the time. (Tried deleting their mailboxes too with no luck- shot in the dark). No Anti-Virus software running on this particular server (I know, -bad idea) so rules that out. Loaded anti-virus software on the machine after the problem occurred and scanned the system hard. Nothing found. I'm working on setting up to do the ProudUSA fix now. Wonder how many other folks are pulling their hair out over this that we don't know about?

Mike

 

[twillboy]

Well, I don't think that we have an AV software installed on our Exchange server (DON'T ASK ME WHY) My suggestion to the Admin was to get something installed ASAP.

2nd - We don't have the same event log errors as Alan, but our event log did go crazy at about 11 am cst on the 19th. I am waiting to get a copy of that event log from the admin so I can analyze those particular logs

(I by the way don't have anything to do with the company that is having the problems... I'm a friend of one of the assistant admins and they asked for my opinion/ideas/solutions/etc I am just trying to help out.)

Anyhow, will update the board with the event log errors that we got.

 

[ALAN1980]

I forgot to mention my workaround for my users (my new drive will not be in until Friday). There are two:

1. Outlook Express -- We use this to collect the user's e-mail from our POP mail server. This is the workaround that worked best for us.

2. Create a new Personal Folder in Outlook -- I do not recommend this options. But if you want try it these are the steps:
-- Create a Personal Folder
-- From Outlook > Tools > Services > Delivery
-- Then change "Deliver new mail to the following locations" to Personal Folders. Make sure to restart Outlook.

WARNING: All your old e-mails will be moved to your Personal Folder inbox automatically by Outlook. Also any appointments from you old calendar will not pop-up as reminders. All new internal and Internet e-mails will be sent to the Personal Folders. We tried it but it created more problems. That is why I we are using the first fix.

 

[twillboy]

Okay, after looking at the event log I have found that it contains nothing really related to the issues at hand. It is all DHCP stuff that went down and was brought back up.

As for hard drive space, the drive is (I believe) about 15 Gb and we have 950+ MB free. I know that everything was working fine before when there was only 700+ MB free so I don't think this is the issue at all.

I did have to get DHCP fixed though. The router ip was removed from the DHCP Manager so all Win95 and 98 boxes were not receiving a gateway nor could one be set manually due to 98 and 95s TCP/IP stack only allowing dynamic allocation or static but not a combination like 2000 and XP where you can specify a static Gateway but still receive a dynamic IP.

Anyhow, that's working again so I'm now back at square one. I don't relish the thought of telling these people to rebuild their machine since this exchange server is home for DNS, DHCP and everything else on this network.

exciting isn't it? I am running out of ideas. Being a small town hospital, they don't have a ton of money budgeted for IT and therefor we can't implement any big-money solutions like buying more servers to split up tasks so that in the event that a rebuild is necessary it won't kill everyone. (funny how that could even be a literal meaning too)

I'm up for more ideas/thoughts/solutions to try.

Thanks.

 

[ALAN1980]

Twillboy I have two questions and one observation:

1. What is the size of your PRIV.EDB & PUB.EDB files?

2. Did you check the Event Viewer -> Application logs?

Observation -- If you only have 950MB free out of a 15GB drive you are in big ... big trouble. Your Event Viewer should show at least a warning that you are running low on drive space. Also, as I discovered, your Exchange files PRIV.EDB & PUB.EDB are not being defrag because if you do not have enough disk space. Which may be a problem. Ideas, I have plenty, but sadly they cost money.

 

[twillboy]

You know Alan, that is something I didn't even think of to check at the time.

I will check both of these things on the server when I get a chance to do so.

As far as ideas that cost money, when it's situations like this, users will probably be asked to throw throw throw or administration will just have to choke up some more money for a new HD on the server. I have a feeling that those 2 ideas are the only options that they are going to be able to go with.

Thank you very much and I will post my findings on those 2 items.

 

[twillboy]

another thought, if this was a problem with space on the server then why does it not cause problems for most of my 2000 and XP users but only approximately 30 of my 98/95 users.

 

[jonarden]

Some additional thoughts:
1. mikewy has the Outlook problem, but doesn't have a drive space issue, so lack of drive space (though an important issue to resolve) is likely not the cause of the Outlook issue.
2. ALAN1980 and twillboy you probably know this, but you can check the Application log for the last complete Event 1221 for both the pub and priv databases to see if any available free space in the databases could be recovered with an offline defrag. You'll have to move the databases you want to defrag to another server (or drive) if you don't have free space = 110% of the database you are defragging on the database's drive. See Knowledge Base articles KB192185 "How to Defragment with the Esutil Utility (Esutil.exe)" & KB244525 "How to run Eseutil on a Computer Without Exchange Server".
3. mikewy and others, there's investigation on another thread of AOL apparently enabling some type of Outlook blocking of POP3 access during the date range your issue occurred. I'll locate & post the thread number in case they come up with info that might also be useful to you.