General discussion

Locked

Users using databases experience crashes

By Eric the Half-a-Bee ·
Background: was running 2 WinNT Server (SP6) machines, one with Exchange 5.5 (SP4). Added a W2K3 server and upgraded the non-Exchange machine to W2K3. Installed Exchange 2K3 on one of the upgraded machines. Created a new AD Domain. Migrated all accounts. Migrated e-mail from Exchange 5.5 to Exchange 2K3. Turned off the only remaining WinNT machine and set it on a shelf for 10 months. So far so good, and everything works fine. Until.

Two months ago I decided to "properly" remove the last server in a domain to get rid of the NT machine for good. Well, I could not actually get the WinNT machine removed because of reported incomplete replication. But everything was working without the machine being connected; I just wanted to clean the servers up a bit. So I scoured the MS (and other) websites and found a variety of ways to get rid of the information, and managed to do so. And that's when trouble began happening.

Before I continue: the DC in the AD Domain is also our file/application server (so various databases are housed there in addition to user files). The other server in the AD domain is a member server, but not a DC. It hosts web pages and e-mail.

Not long after I had obliterated the last traces of the former domain, several things happened. People using ACT suddenly started experiencing crashes. Same with people using Access (one person). Same with the lone GroupWise holdout. Same with the development office software and the accounting office software. Most users had no problems. Since when the crashes initially happened each person reported that they had Outlook open, I began investigating the possibility that the problem lie somewhere in having removed the Exchange 5.5 server (unfortunately, closing Outlook did not stop the crashing).

How likely is it that the databases crashing is related to improperly removing the now non-existent domain and its controller? How likely is it that the problem is network-related and simply a m

This conversation is currently closed to new comments.

7 total posts (Page 1 of 1)  
| Thread display: Collapse - | Expand +

All Comments

Collapse -

by Eric the Half-a-Bee In reply to Users using databases exp ...

Let me finish that last sentence:

"How likely is it that the databases crashing is related to improperly removing the now non-existent domain and its controller? How likely is it that the problem is network-related and simply a matter of horrible timing (I'm currently investigating this)?"

Collapse -

by Eric the Half-a-Bee In reply to Users using databases exp ...

A bit more information:

I've run netdiag and dcdiag (on the DC of course), and everything checks out. I've run ntdsutil and done metadata cleanup, I've run offline defragmentation and error checking on the AD database, I've attempted some things with ldp ... though I don't recall anymore what exactly, and I've deleted and recreated the DNS structure. I have not yet recreated the entire domain as that will result, if I understand it correctly, in my having to recreate all sorts of other things like shares and users. So I'm reticent, but will do this if no other suggestions yield results.

When running ntdsutil against the member server (non-DC) and attempt to connect to it I get a DsBindW (0x6d9) no RPC endpoints available from the endpoint mapper, but fail to see why this particular server even enters into the picture.

I also turned up LDAP logging in the event log and am receiving two NTDS LDAP events (both labeled informational). The first is from the member server:

USER: DOMAIN\MACHINE$
Internal event: The LDAP server returned an error.

Additional Data
Error value:
000020: NameErr: DSID-031001A8, problem 2001 (NO_OBJECT), data 0, best match of:
''

The second is similar, but from the System account:

USER: NT AUTHORITY\SYSTEM
Internal event: The LDAP server returned an error.

Additional Data
Error value:
000020: NameErr: DSID-031001CD, problem 2001 (NO_OBJECT), data 0, best match of:
'CN=Dfs-Configuration,CN=System,DC=my,DC=domain,DC=com'

This seems likely to be related, but I don't actually know with certainty since I had no reason to increase LDAP logging prior to attempting to solve this little beauty.

Collapse -

by Eric the Half-a-Bee In reply to Users using databases exp ...

Hey, I have no idea how much this is worth in the grand scheme of things. I'll give away all my points if I have to.

Collapse -

by sgt_shultz In reply to Users using databases exp ...

please accept my sympathies. without knowing the steps you took to um remove the last traces of the last winNT svr box (i'm sorry, what did you say its function(s) were exactly?
don't think can guess if related. i hate it when timing like that happens...
please tell more about the crashes. what happens? lock up? how do they recover. reboot? close app and reopen? do they lose data? does it get corrupted. can you get them to keep a log. with time and date and how they recovered and any error message you get. can post exact error id, source and description if any on bad clients
can you reproduce this at will (you can't, right?)
what else has changed?
sp1?
new hd controller/driver?
what hardware do the databases live on.

Collapse -

by Eric the Half-a-Bee In reply to

Ah, yes, sgt_shultz -- I guess some of that info would have been helpful ...

As much as possible I followed MS Q272314 for removing the last Exchange 5.5 server from a site. There were some things that did not work (though it's been a while, and I'm having difficulty recalling what, exactly), for which I found other answers. Seriously, there was a pile of searching that took place and discomfortingly little information.

The box I attempted to remove was the sole domain controller and e-mail server on an obsolete domain (we moved from NT on one domain to a new domain on 2K3). User accounts and e-mail seemed to be working fine without benefit of the old box -- that is, everything seemed to have migrated successfully.

The crashes differ with different apps. Nothing shows in the event log on the server, nor, so far as I've been able to determine, on the client. I've also turned up user environment logging but that does not reveal much of worth to me. Some apps indicate ODBC troubles; some "too many windows open"; some simply say "Access is denied." And users don't necessarily all crash at once, though sometimes they have. I have, of course, sought information regarding each error message, but none of the proposed solutions solve the problem.

I'd be a happy camper if I could reproduce this at will. But I can't.

2K3 SP1, Exchange 2K3 SP1. The Windows SP I installed hoping it would resolve the problem, but it did not. I also updated to 2000 Native Mode (but not 2K3 Native Mode). Otherwise nothing new. Server is Dell PowerEdge 2600. It's occurred to me that perhaps a controller is flaky, but I've not yet tried anything to address that.

Application recovery by restarting application in most cases; sometimes rebooting (but that, I believe, could be avoided by judiciously killing an orphaned process).

Collapse -

by montelski In reply to Users using databases exp ...

Don't want any points, Just checking to make sure you didn't overlook pointing the clients DNS to the new 2K3 server.(common oversight) I failed to do this and started having random crashes and disconnects from the server. Changed the primary DNS of the clients to the 2K3 server and fixed my problem.

Collapse -

by Eric the Half-a-Bee In reply to

DNS is handled via DHCP. The good news is that today the server started crashing. Which points to a hardware problem.

Back to Networks Forum
7 total posts (Page 1 of 1)  

Related Discussions

Related Forums