Showing posts with label servers. Show all posts
Showing posts with label servers. Show all posts

Thursday, March 29, 2012

Can I Have Link Between Several servers

hello all

I have a web application(asp.net) and Database(sql server 2005) .

we have installed them on several servers.

now we want to have a connection between apllication on one server to a database on another server .

for example we have Server(A) and Server(B) the DataBases and Web apllications on two server are the same.

sometimes the Apllication on server(A) must connect to database on server(B).

whats the solution plz?

note that the number of servers can be inceased in the future this mean the number of servers are not fixed.

thanks

Use the function sp_linkedServer of the SQL to link the borh servers.

helped?

Regards

Can I Have link Between Several Servers

hello all

I have a web application(asp.net) and Database(sql server 2005) .

we have installed them on several servers.

now we want to have a connection between apllication on one server to a database on another server .

for example we have Server(A) and Server(B) the DataBases and Web apllications on two server are the same.

sometimes the Apllication on server(A) must connect to database on server(B).

whats the solution plz?

note that the number of servers can be inceased in the future this mean the number of servers are not fixed.

thanks

Not sure I follow -- do you mean you have 2 different database serves that should mirror each other, ie, the data is the same and updates on one should be mirrored on the other?

|||

You have a database in the server A .. connect to it by using ConnectionStringA , and any need data from Server B .. use another ConnectionString (e.g. ConnectionStringB).

So, in the web.config file you will have tow ConnectionStrings (A & B).. when to connect to the database in the same server use StringConnectionA and the other used to connect to the database in the another server (server B) - called ConnectionString B.

Good luck.

|||

thank you

for Contributor : i have defrerent data in each databas i cant use replication

for Participant : i must connect to different web server .

i get it that i must to use splinkserver to link servers

thanks forever.

Tuesday, March 27, 2012

Can I ghost an SQL server?

I have a test environment that we rebuild servers on a regular basis.
To streamline the process we use ghost. We will be installing SQL on
the servers and want to build a ghost image with that build. We have
tested it by doing the build, loading the data, and then stopping all
the services and setting them to manual. After we Ghost the machine,
we start up the services and reset them to Automatic. Seems to work.

My question is:

Are there any risks? Should I expect any adverse affects?

Thanks for all your help!(otisim@.YAHOO.COM) writes:
> I have a test environment that we rebuild servers on a regular basis.
> To streamline the process we use ghost. We will be installing SQL on
> the servers and want to build a ghost image with that build. We have
> tested it by doing the build, loading the data, and then stopping all
> the services and setting them to manual. After we Ghost the machine,
> we start up the services and reset them to Automatic. Seems to work.
> My question is:
> Are there any risks? Should I expect any adverse affects?

Well, there is one catch. I assume that you rename the machines once
the ghosting as been completed. In this case @.@.servername is likely
to retain the old name or be NULL. In the latter case, you need to do:

sp_addserver NEWNAME, local

Restart SQL Server after this. If @.@.servername retains the old name
I would try dropping it with sp_dropserver first, although I don't
know if it works.

--
Erland Sommarskog, SQL Server MVP, esquel@.sommarskog.se

Books Online for SQL Server SP3 at
http://www.microsoft.com/sql/techin.../2000/books.asp|||Actually, we are keeping the same name. This is because it is just a
rebuild of the same server.

In other words, we do some testing, we reimage the machine (so we start
with a fresh standard image), and then do more testing.

So we want to keep the same name. We are not using sysprep or making
any other changes. Same hardware, same build, same everything. Just
want to clear out any changes that were made during testing.

Thanks so much for the response.

Steve

Thursday, March 22, 2012

Can I Created a BPA 'Job'

Can I created a job that I can send to my Service Delivery group to run on the Production servers? I'm a developer and don't WANT access to PROD, but I want BPA run against PROD. I'd like to have them install BPA and them send them a 'package' to run. Is
this available with BPA?
Thanks,
Maybe not exactly what you want, but there's a way you can get close.
Install BPA, which creates the repository. In it, create a best practice
group that scans things the way you want. You should register the servers
from production, or register them as (local).
Have your production team install BPA (they will have to create a repository
but it can be discarded) and then ask them to run bpa (likely the command
line version bpacmd.exe) pointing at your repository database. If they do
that, they'll have to execute the best practice group that you created.
- Christian
"B.W." <bw.lightsey@.tetrapak.com> wrote in message
news:611E0DF9-C69E-4EF6-AA6B-0E390C3430A8@.microsoft.com...
> Can I created a job that I can send to my Service Delivery group to run on
the Production servers? I'm a developer and don't WANT access to PROD, but I
want BPA run against PROD. I'd like to have them install BPA and them send
them a 'package' to run. Is this available with BPA?
> Thanks,
>
sql

Sunday, March 11, 2012

Can I access MSDE on a remote machine with Enterprise Manager?

We just installed an application, WhatsUpGold, to monitor our servers. Management won't let me host that on my primary SQL box. So, we are using MSDE on the local machine (call that box my_netmon).

I can access the db with very basic functionality using osql while on the my_netmon machine. I've added my domain user account as a System Admin using the following commands.

exec sp_grantlogin @.name
go

exec sp_addsrvrolemember @.name, 'sysadmin'

But when I go to the my_sql machine and try to attach to it with Enterprise Manager I get an error that the server does not exist.

What gives? Is it just not possible, or am I missing something here?

My goal is to just be able to see the data stored in the MSDE db. Write views, etc...

Please advise,

MarcHas anybody accessed MSDE from a remote machine. Or can it only be access locally with ODBC type connections?|||yes, you can use MSDE remotely. what does your connection string look like? is it a named instance?

for that matter, why bother with MSDE? it's ancient. you should use SQL Server Express if this is for new development.

Can default and named instance co-exist?

I have 2 Windows 2003 servers running SQL2000. They each have separate
databases. I am in the process of clustering these 2 machines together. I
have purchased a shared external SCSI PowerVault appliance to put the quorom
and the shared databases on. I want to create a Virtual SQL server using
these 2 machines and put their separate databases on the virtual server. My
question is:
Do I have to uninstall SQL server on each machine before I install the
virtual server? If I have to I would have to restore my databases to the new
virtual server
Or can I just install a new named instance of SQL server - choosing the
virtual server in the setup? If I can do this - it would be a lot easier
moving my databases over to the new server.
If I can just create a new virtual server with leaving the original SQL in
tact - will the original still be available to my users if I decide not to
move all the databases over that day?
Is there one way that is better than the other?
1. If you want to use the exact same instance names, uninstall/reinstall for
clustering is needed. If not, you can just virtualize your new instance and
migrate data.
2. Consider reading through this kb for some details
http://support.microsoft.com/kb/224071
3. Your original instance will continue to run while you're installing a new
instance on the same node.
The bottom line, I would create a new instance and ensure it's up and
running before tinkering with the original.
-oj
"Amy Lewis" <AmyLewis@.discussions.microsoft.com> wrote in message
news:223F3934-2F36-4A2D-B5D5-D860055AE5E7@.microsoft.com...
>I have 2 Windows 2003 servers running SQL2000. They each have separate
> databases. I am in the process of clustering these 2 machines together.
> I
> have purchased a shared external SCSI PowerVault appliance to put the
> quorom
> and the shared databases on. I want to create a Virtual SQL server using
> these 2 machines and put their separate databases on the virtual server.
> My
> question is:
> Do I have to uninstall SQL server on each machine before I install the
> virtual server? If I have to I would have to restore my databases to the
> new
> virtual server
> Or can I just install a new named instance of SQL server - choosing the
> virtual server in the setup? If I can do this - it would be a lot easier
> moving my databases over to the new server.
> If I can just create a new virtual server with leaving the original SQL in
> tact - will the original still be available to my users if I decide not to
> move all the databases over that day?
> Is there one way that is better than the other?
|||Each instance of SQL Server requires its own shared disk. You can install
either 1 default instance and 15 named instances or up to 16 named
instances, but each must have its own shared drive. All drives must have a
drive letter, no mount points allowed. For existing nonclustered instances,
if you move the data files to a shared drive then you will have the option
during setup to upgrade the standalone instance to a virtual/clustered
instance (assuming you're using the Enterprise Edition setup).
Cindy Gross, MCDBA, MCSE
http://cindygross.tripod.com
This posting is provided "AS IS" with no warranties, and confers no rights.

Wednesday, March 7, 2012

can connect to all servers but one on certain box

I have an issue where users can't connect to a server - Server A - from a pa
rticular client computer - Box B. They can connect to other servers from Box
B - even servers in the same cluster as Server A. On other client machines
the same users can connect
to all servers - including Server A. They are using Windows authorization. I
s there any setting that might need to be changed either in Windows or SQL S
erver on either Server A or Box B that might correct this situation?Sounds like you want to check Client Network Utility and Server Network
Utility on the involved machines.
Tibor Karaszi, SQL Server MVP
http://www.karaszi.com/sqlserver/default.asp
"hawthorne" <anonymous@.discussions.microsoft.com> wrote in message
news:A3F01255-54EF-4F09-A8AD-0CBA21C6D04B@.microsoft.com...
> I have an issue where users can't connect to a server - Server A - from a
particular client computer - Box B. They can connect to other servers from
Box B - even servers in the same cluster as Server A. On other client
machines the same users can connect to all servers - including Server A.
They are using Windows authorization. Is there any setting that might need
to be changed either in Windows or SQL Server on either Server A or Box B
that might correct this situation?|||Can you PING Server A from Box B ?

CAN CONNECT ONE WAY TO SQL SERVER BUT NOT BACK

There is a Remote SQL server 2000 server that can connect (register
thru query analyzer or enterprise manager) with 3 remote Sql server
2000 servers without issue. When these 3 servers try and connect back
(thru query analyzer or enterprise manager) I get the "SQL Server does
not exist or access denied" error message.
I have confirmed with my network guys that the request is making it to
the server so the failure point is at the server (Windows 2003). the
firewall is turned off so I know that is not the issue.
Any insight would be very helpful.
you may read the following kb for a possible resolution:
http://support.microsoft.com/default...b;en-us;827422
-billyy
"csmftbb" <andrew.farris@.sbcglobal-dot-net.no-spam.invalid> ?
news:e5qdnXI6DIJiKgvfRVn_vg@.giganews.com...
> There is a Remote SQL server 2000 server that can connect (register
> thru query analyzer or enterprise manager) with 3 remote Sql server
> 2000 servers without issue. When these 3 servers try and connect back
> (thru query analyzer or enterprise manager) I get the "SQL Server does
> not exist or access denied" error message.
> I have confirmed with my network guys that the request is making it to
> the server so the failure point is at the server (Windows 2003). the
> firewall is turned off so I know that is not the issue.
> Any insight would be very helpful.
>

CAN CONNECT ONE WAY TO SQL SERVER BUT NOT BACK

There is a Remote SQL server 2000 server that can connect (register
thru query analyzer or enterprise manager) with 3 remote Sql server
2000 servers without issue. When these 3 servers try and connect back
(thru query analyzer or enterprise manager) I get the "SQL Server does
not exist or access denied" error message.
I have confirmed with my network guys that the request is making it to
the server so the failure point is at the server (Windows 2003). the
firewall is turned off so I know that is not the issue.
Any insight would be very helpful.you may read the following kb for a possible resolution:
http://support.microsoft.com/defaul...kb;en-us;827422
-billyy
"csmftbb" <andrew.farris@.sbcglobal-dot-net.no-spam.invalid> '
news:e5qdnXI6DIJiKgvfRVn_vg@.giganews.com...
> There is a Remote SQL server 2000 server that can connect (register
> thru query analyzer or enterprise manager) with 3 remote Sql server
> 2000 servers without issue. When these 3 servers try and connect back
> (thru query analyzer or enterprise manager) I get the "SQL Server does
> not exist or access denied" error message.
> I have confirmed with my network guys that the request is making it to
> the server so the failure point is at the server (Windows 2003). the
> firewall is turned off so I know that is not the issue.
> Any insight would be very helpful.
>

Saturday, February 25, 2012

Can anyone help solve this puzzle?

I have one SQL instance that is giving me problems: everytime I restore a
database to it, the database has many consistency errors.
I have 3 servers I can test with: Server A, Server B, and Server C. Server
C is the one that's giving me problems.
I run a DBCC CheckDB on a database on Server A, then back it up. If I then
try to restore that backup on Server C, the backup works but DBCC CheckDB
reveals numerous errors. However, I can take that same backup file and
restore it on Server B, and everything is fine.
I can spend lots of time with my corrupt database on Server C and using DBCC
CheckDB (repair_rebuild or repair_allow_data_loss) and dbreindex to finally
get the database to report that it has no consitency errors. HOWEVER, if I
then try to back up that database and restore it on either Server A or
Server B, the restore works ok, but the database again is full of
consistency errors. Restore/backup works with no problems whatsoever
between Server A and Server B.
All of these problems apply not only to backup/restore, but also if I try to
just detach from Server A (or B) and attach to Server C.
I have sees several errors in the SQL log, on Server C, 604 (and some
others). I wish I had the complete list but I have re-installed SQL and
lost the logs.
I thought maybe it was an problem with the install, but as I just mentioned,
I completely re-installed on Server C, but the problems continue. To this
point in the current logs, there are no errors.
At this point I am thinking hardware, but the network guys insist that it
has passed all tests. Next I am going to run SQLIOStress (
http://support.microsoft.com/default.aspx?scid=http://support.microsoft.com:80/support/kb/articles/q231/6/19.asp&NoWebContent=1 )
and see if that turns anything up.
Oh, info on the servers:
Server A: SQL: 8.00.194 RTM Enterprise Edition OS: Windows 2000
(Build 2195 SP3)
Server B: SQL: 8.00.760 SP3 Developer Edition OS: Windows Server 2003
Standard Edition (Build 3790.srv03_rtm.030324-2048)
Server C: SQL: 8.00.760 SP3 Enterprise Edition OS: Windows 2000
(Build 2195 SP4) NOTE: the new install has no SQL service packs applied, no
difference in behavior
Any help/suggestions/links would be greatly appreciated. Please post back
here so that maybe any other unlucky soul that has this problem can benefit
from the smart folks out there.
thanks,
Greg
gjleclair AT hotmail DOT comGreg,
Step 1: install SQL Server service pack 3a and Win2K service pack 4 on
server A.
Step 2: see if you are still having trouble.
I wouldn't be so sure it's server C with the problem. Server A, which
has no SQL Server service packs installed, created the backup.
You mentioned briefly a problem detaching from B and attaching to C. If
that is a problem that is reproducible from scratch without any contact
with Server A, post the details of what's going on with that.
There are hardware issues that can cause this kind of problem that might
not be called hardware problems, such as the one described in
http://support.microsoft.com:80/support/kb/articles/q268/4/81.asp, but I
doubt you'll get much help asking why things aren't working on an
non-updated server.
SK
Greg wrote:
>I have one SQL instance that is giving me problems: everytime I restore a
>database to it, the database has many consistency errors.
>I have 3 servers I can test with: Server A, Server B, and Server C. Server
>C is the one that's giving me problems.
>I run a DBCC CheckDB on a database on Server A, then back it up. If I then
>try to restore that backup on Server C, the backup works but DBCC CheckDB
>reveals numerous errors. However, I can take that same backup file and
>restore it on Server B, and everything is fine.
>I can spend lots of time with my corrupt database on Server C and using DBCC
>CheckDB (repair_rebuild or repair_allow_data_loss) and dbreindex to finally
>get the database to report that it has no consitency errors. HOWEVER, if I
>then try to back up that database and restore it on either Server A or
>Server B, the restore works ok, but the database again is full of
>consistency errors. Restore/backup works with no problems whatsoever
>between Server A and Server B.
>All of these problems apply not only to backup/restore, but also if I try to
>just detach from Server A (or B) and attach to Server C.
>I have sees several errors in the SQL log, on Server C, 604 (and some
>others). I wish I had the complete list but I have re-installed SQL and
>lost the logs.
>I thought maybe it was an problem with the install, but as I just mentioned,
>I completely re-installed on Server C, but the problems continue. To this
>point in the current logs, there are no errors.
>At this point I am thinking hardware, but the network guys insist that it
>has passed all tests. Next I am going to run SQLIOStress (
>http://support.microsoft.com/default.aspx?scid=http://support.microsoft.com:80/support/kb/articles/q231/6/19.asp&NoWebContent=1 )
>and see if that turns anything up.
>
>Oh, info on the servers:
>Server A: SQL: 8.00.194 RTM Enterprise Edition OS: Windows 2000
>(Build 2195 SP3)
>Server B: SQL: 8.00.760 SP3 Developer Edition OS: Windows Server 2003
>Standard Edition (Build 3790.srv03_rtm.030324-2048)
>Server C: SQL: 8.00.760 SP3 Enterprise Edition OS: Windows 2000
>(Build 2195 SP4) NOTE: the new install has no SQL service packs applied, no
>difference in behavior
>
>Any help/suggestions/links would be greatly appreciated. Please post back
>here so that maybe any other unlucky soul that has this problem can benefit
>from the smart folks out there.
>
>thanks,
>Greg
>gjleclair AT hotmail DOT com
>
>|||Thanks for the reply, Steve.
I understand that applying the service packs would make the most sense as a
starting point...but the powers that be do not want me to do that since
"that's only one that's working".
The reason I didn't think it was server A is that all interaction between
Server A and Server B (backup/restore, attach/detach) work fine. Any
interaction between Server B and Server C result in the same consistency
problems.
As an update the SQLIOstress failed out with a series of these errors:
>>>>*** ERROR: LSN for page 277534 is out of sequence. Expected:
605885 and found 601789 in sector 15. Probably a torn page.
>>>> Byte 513 in pattern is [>] and from read is [.]
Followed by this:
>>>>----
--
>>>>Expected pattern [C] in file for page 277534.
>>>>Bytes read = 8192
>>>>Potential torn write, lost write or stale read may have been
encountered
>>>>----
--
I'm guessing that we have a hardware issue. Any thoughts? It seems like it
is up to me to "prove" that the hardware is having issues but I'm not really
sure what those errors indicate.
Once again, any help is appreciated.
-Greg
"Steve Kass" <skass@.drew.edu> wrote in message
news:%23bYZld26DHA.2556@.TK2MSFTNGP09.phx.gbl...
> Greg,
> Step 1: install SQL Server service pack 3a and Win2K service pack 4 on
> server A.
> Step 2: see if you are still having trouble.
> I wouldn't be so sure it's server C with the problem. Server A, which
> has no SQL Server service packs installed, created the backup.
> You mentioned briefly a problem detaching from B and attaching to C. If
> that is a problem that is reproducible from scratch without any contact
> with Server A, post the details of what's going on with that.
> There are hardware issues that can cause this kind of problem that might
> not be called hardware problems, such as the one described in
> http://support.microsoft.com:80/support/kb/articles/q268/4/81.asp, but I
> doubt you'll get much help asking why things aren't working on an
> non-updated server.
> SK
> Greg wrote:
> >I have one SQL instance that is giving me problems: everytime I restore
a
> >database to it, the database has many consistency errors.
> >I have 3 servers I can test with: Server A, Server B, and Server C.
Server
> >C is the one that's giving me problems.
> >
> >I run a DBCC CheckDB on a database on Server A, then back it up. If I
then
> >try to restore that backup on Server C, the backup works but DBCC CheckDB
> >reveals numerous errors. However, I can take that same backup file and
> >restore it on Server B, and everything is fine.
> >
> >I can spend lots of time with my corrupt database on Server C and using
DBCC
> >CheckDB (repair_rebuild or repair_allow_data_loss) and dbreindex to
finally
> >get the database to report that it has no consitency errors. HOWEVER, if
I
> >then try to back up that database and restore it on either Server A or
> >Server B, the restore works ok, but the database again is full of
> >consistency errors. Restore/backup works with no problems whatsoever
> >between Server A and Server B.
> >
> >All of these problems apply not only to backup/restore, but also if I try
to
> >just detach from Server A (or B) and attach to Server C.
> >
> >I have sees several errors in the SQL log, on Server C, 604 (and some
> >others). I wish I had the complete list but I have re-installed SQL and
> >lost the logs.
> >
> >I thought maybe it was an problem with the install, but as I just
mentioned,
> >I completely re-installed on Server C, but the problems continue. To
this
> >point in the current logs, there are no errors.
> >
> >At this point I am thinking hardware, but the network guys insist that it
> >has passed all tests. Next I am going to run SQLIOStress (
>http://support.microsoft.com/default.aspx?scid=http://support.microsoft.com
:80/support/kb/articles/q231/6/19.asp&NoWebContent=1 )
> >and see if that turns anything up.
> >
> >
> >Oh, info on the servers:
> >
> >Server A: SQL: 8.00.194 RTM Enterprise Edition OS: Windows 2000
> >(Build 2195 SP3)
> >Server B: SQL: 8.00.760 SP3 Developer Edition OS: Windows Server
2003
> >Standard Edition (Build 3790.srv03_rtm.030324-2048)
> >Server C: SQL: 8.00.760 SP3 Enterprise Edition OS: Windows 2000
> >(Build 2195 SP4) NOTE: the new install has no SQL service packs applied,
no
> >difference in behavior
> >
> >
> >Any help/suggestions/links would be greatly appreciated. Please post
back
> >here so that maybe any other unlucky soul that has this problem can
benefit
> >from the smart folks out there.
> >
> >
> >thanks,
> >Greg
> >
> >gjleclair AT hotmail DOT com
> >
> >
> >
> >
>|||Greg,
It sure sounds like a hardware issue, but I don't have any great
suggestions. If you haven't already looked, see if these Knowledge Base
articles help at all:
http://support.microsoft.com:80/support/kb/articles/q231/6/19.asp
(and the articles it references at the bottom)
http://support.microsoft.com/default.aspx?scid=kb;en-us;268481
It is sounding more like it's server C and not the service pack, but
you never know... It does seem a little strange that restore fails so
easily, yet you have to work to get SQLIOstress to error out.
SK
Greg wrote:
>Thanks for the reply, Steve.
>I understand that applying the service packs would make the most sense as a
>starting point...but the powers that be do not want me to do that since
>"that's only one that's working".
>The reason I didn't think it was server A is that all interaction between
>Server A and Server B (backup/restore, attach/detach) work fine. Any
>interaction between Server B and Server C result in the same consistency
>problems.
>As an update the SQLIOstress failed out with a series of these errors:
>
>
>>>>*** ERROR: LSN for page 277534 is out of sequence. Expected:
>>>>
>>>>
>605885 and found 601789 in sector 15. Probably a torn page.
>
>>>>Byte 513 in pattern is [>] and from read is [.]
>>>>
>>>>
>Followed by this:
>
>>>>----
>>>>
>>>>
>--
>
>>>>Expected pattern [C] in file for page 277534.
>>>>Bytes read = 8192
>>>>Potential torn write, lost write or stale read may have been
>>>>
>>>>
>encountered
>
>>>>----
>>>>
>>>>
>--
>I'm guessing that we have a hardware issue. Any thoughts? It seems like it
>is up to me to "prove" that the hardware is having issues but I'm not really
>sure what those errors indicate.
>Once again, any help is appreciated.
>-Greg
>
>
>"Steve Kass" <skass@.drew.edu> wrote in message
>news:%23bYZld26DHA.2556@.TK2MSFTNGP09.phx.gbl...
>
>>Greg,
>>Step 1: install SQL Server service pack 3a and Win2K service pack 4 on
>>server A.
>>Step 2: see if you are still having trouble.
>>I wouldn't be so sure it's server C with the problem. Server A, which
>>has no SQL Server service packs installed, created the backup.
>>You mentioned briefly a problem detaching from B and attaching to C. If
>>that is a problem that is reproducible from scratch without any contact
>>with Server A, post the details of what's going on with that.
>>There are hardware issues that can cause this kind of problem that might
>>not be called hardware problems, such as the one described in
>>http://support.microsoft.com:80/support/kb/articles/q268/4/81.asp, but I
>>doubt you'll get much help asking why things aren't working on an
>>non-updated server.
>>SK
>>Greg wrote:
>>
>>I have one SQL instance that is giving me problems: everytime I restore
>>
>a
>
>>database to it, the database has many consistency errors.
>>I have 3 servers I can test with: Server A, Server B, and Server C.
>>
>Server
>
>>C is the one that's giving me problems.
>>I run a DBCC CheckDB on a database on Server A, then back it up. If I
>>
>then
>
>>try to restore that backup on Server C, the backup works but DBCC CheckDB
>>reveals numerous errors. However, I can take that same backup file and
>>restore it on Server B, and everything is fine.
>>I can spend lots of time with my corrupt database on Server C and using
>>
>DBCC
>
>>CheckDB (repair_rebuild or repair_allow_data_loss) and dbreindex to
>>
>finally
>
>>get the database to report that it has no consitency errors. HOWEVER, if
>>
>I
>
>>then try to back up that database and restore it on either Server A or
>>Server B, the restore works ok, but the database again is full of
>>consistency errors. Restore/backup works with no problems whatsoever
>>between Server A and Server B.
>>All of these problems apply not only to backup/restore, but also if I try
>>
>to
>
>>just detach from Server A (or B) and attach to Server C.
>>I have sees several errors in the SQL log, on Server C, 604 (and some
>>others). I wish I had the complete list but I have re-installed SQL and
>>lost the logs.
>>I thought maybe it was an problem with the install, but as I just
>>
>mentioned,
>
>>I completely re-installed on Server C, but the problems continue. To
>>
>this
>
>>point in the current logs, there are no errors.
>>At this point I am thinking hardware, but the network guys insist that it
>>has passed all tests. Next I am going to run SQLIOStress (
>>
>>http://support.microsoft.com/default.aspx?scid=http://support.microsoft.com
>>
>:80/support/kb/articles/q231/6/19.asp&NoWebContent=1 )
>
>>and see if that turns anything up.
>>
>>Oh, info on the servers:
>>Server A: SQL: 8.00.194 RTM Enterprise Edition OS: Windows 2000
>>(Build 2195 SP3)
>>Server B: SQL: 8.00.760 SP3 Developer Edition OS: Windows Server
>>
>2003
>
>>Standard Edition (Build 3790.srv03_rtm.030324-2048)
>>Server C: SQL: 8.00.760 SP3 Enterprise Edition OS: Windows 2000
>>(Build 2195 SP4) NOTE: the new install has no SQL service packs applied,
>>
>no
>
>>difference in behavior
>>
>>Any help/suggestions/links would be greatly appreciated. Please post
>>
>back
>
>>here so that maybe any other unlucky soul that has this problem can
>>
>benefit
>
>>from the smart folks out there.
>>
>>thanks,
>>Greg
>>gjleclair AT hotmail DOT com
>>
>>
>>
>
>

Can anyone help solve this puzzle?

I have one SQL instance that is giving me problems: everytime I restore a
database to it, the database has many consistency errors.
I have 3 servers I can test with: Server A, Server B, and Server C. Server
C is the one that's giving me problems.
I run a DBCC CheckDB on a database on Server A, then back it up. If I then
try to restore that backup on Server C, the backup works but DBCC CheckDB
reveals numerous errors. However, I can take that same backup file and
restore it on Server B, and everything is fine.
I can spend lots of time with my corrupt database on Server C and using DBCC
CheckDB (repair_rebuild or repair_allow_data_loss) and dbreindex to finally
get the database to report that it has no consitency errors. HOWEVER, if I
then try to back up that database and restore it on either Server A or
Server B, the restore works ok, but the database again is full of
consistency errors. Restore/backup works with no problems whatsoever
between Server A and Server B.
All of these problems apply not only to backup/restore, but also if I try to
just detach from Server A (or B) and attach to Server C.
I have sees several errors in the SQL log, on Server C, 604 (and some
others). I wish I had the complete list but I have re-installed SQL and
lost the logs.
I thought maybe it was an problem with the install, but as I just mentioned,
I completely re-installed on Server C, but the problems continue. To this
point in the current logs, there are no errors.
At this point I am thinking hardware, but the network guys insist that it
has passed all tests. Next I am going to run SQLIOStress (
http://support.microsoft.com/defaul...&NoWebContent=1 )
and see if that turns anything up.
Oh, info on the servers:
Server A: SQL: 8.00.194 RTM Enterprise Edition OS: Windows 2000
(Build 2195 SP3)
Server B: SQL: 8.00.760 SP3 Developer Edition OS: Windows Server 2003
Standard Edition (Build 3790.srv03_rtm.030324-2048)
Server C: SQL: 8.00.760 SP3 Enterprise Edition OS: Windows 2000
(Build 2195 SP4) NOTE: the new install has no SQL service packs applied, no
difference in behavior
Any help/suggestions/links would be greatly appreciated. Please post back
here so that maybe any other unlucky soul that has this problem can benefit
from the smart folks out there.
thanks,
Greg
gjleclair AT hotmail DOT comGreg,
Step 1: install SQL Server service pack 3a and Win2K service pack 4 on
server A.
Step 2: see if you are still having trouble.
I wouldn't be so sure it's server C with the problem. Server A, which
has no SQL Server service packs installed, created the backup.
You mentioned briefly a problem detaching from B and attaching to C. If
that is a problem that is reproducible from scratch without any contact
with Server A, post the details of what's going on with that.
There are hardware issues that can cause this kind of problem that might
not be called hardware problems, such as the one described in
http://support.microsoft.com:80/sup.../q268/4/81.asp, but I
doubt you'll get much help asking why things aren't working on an
non-updated server.
SK
Greg wrote:
quote:

>I have one SQL instance that is giving me problems: everytime I restore a
>database to it, the database has many consistency errors.
>I have 3 servers I can test with: Server A, Server B, and Server C. Server
>C is the one that's giving me problems.
>I run a DBCC CheckDB on a database on Server A, then back it up. If I then
>try to restore that backup on Server C, the backup works but DBCC CheckDB
>reveals numerous errors. However, I can take that same backup file and
>restore it on Server B, and everything is fine.
>I can spend lots of time with my corrupt database on Server C and using DBC
C
>CheckDB (repair_rebuild or repair_allow_data_loss) and dbreindex to finally
>get the database to report that it has no consitency errors. HOWEVER, if I
>then try to back up that database and restore it on either Server A or
>Server B, the restore works ok, but the database again is full of
>consistency errors. Restore/backup works with no problems whatsoever
>between Server A and Server B.
>All of these problems apply not only to backup/restore, but also if I try t
o
>just detach from Server A (or B) and attach to Server C.
>I have sees several errors in the SQL log, on Server C, 604 (and some
>others). I wish I had the complete list but I have re-installed SQL and
>lost the logs.
>I thought maybe it was an problem with the install, but as I just mentioned
,
>I completely re-installed on Server C, but the problems continue. To this
>point in the current logs, there are no errors.
>At this point I am thinking hardware, but the network guys insist that it
>has passed all tests. Next I am going to run SQLIOStress (
>http://support.microsoft.com/defaul...&NoWebContent=1 )
>and see if that turns anything up.
>
>Oh, info on the servers:
>Server A: SQL: 8.00.194 RTM Enterprise Edition OS: Windows 2000
>(Build 2195 SP3)
>Server B: SQL: 8.00.760 SP3 Developer Edition OS: Windows Server 200
3
>Standard Edition (Build 3790.srv03_rtm.030324-2048)
>Server C: SQL: 8.00.760 SP3 Enterprise Edition OS: Windows 2000
>(Build 2195 SP4) NOTE: the new install has no SQL service packs applied, n
o
>difference in behavior
>
>Any help/suggestions/links would be greatly appreciated. Please post back
>here so that maybe any other unlucky soul that has this problem can benefit
>from the smart folks out there.
>
>thanks,
>Greg
>gjleclair AT hotmail DOT com
>
>
|||Thanks for the reply, Steve.
I understand that applying the service packs would make the most sense as a
starting point...but the powers that be do not want me to do that since
"that's only one that's working".
The reason I didn't think it was server A is that all interaction between
Server A and Server B (backup/restore, attach/detach) work fine. Any
interaction between Server B and Server C result in the same consistency
problems.
As an update the SQLIOstress failed out with a series of these errors:
quote:

605885 and found 601789 in sector 15. Probably a torn page.[QUOTE]
Followed by this:
[QUOTE]
--[QUOTE]
encountered[QUOTE]
--
I'm guessing that we have a hardware issue. Any thoughts? It seems like it
is up to me to "prove" that the hardware is having issues but I'm not really
sure what those errors indicate.
Once again, any help is appreciated.
-Greg
"Steve Kass" <skass@.drew.edu> wrote in message
news:%23bYZld26DHA.2556@.TK2MSFTNGP09.phx.gbl...[QUOTE]
> Greg,
> Step 1: install SQL Server service pack 3a and Win2K service pack 4 on
> server A.
> Step 2: see if you are still having trouble.
> I wouldn't be so sure it's server C with the problem. Server A, which
> has no SQL Server service packs installed, created the backup.
> You mentioned briefly a problem detaching from B and attaching to C. If
> that is a problem that is reproducible from scratch without any contact
> with Server A, post the details of what's going on with that.
> There are hardware issues that can cause this kind of problem that might
> not be called hardware problems, such as the one described in
> http://support.microsoft.com:80/sup.../q268/4/81.asp, but I
> doubt you'll get much help asking why things aren't working on an
> non-updated server.
> SK
> Greg wrote:
>
a[QUOTE]
Server[QUOTE]
then[QUOTE]
DBCC[QUOTE]
finally[QUOTE]
I[QUOTE]
to[QUOTE]
mentioned,[QUOTE]
this[QUOTE]
>http://support.microsoft.com/defaul...t.microsoft.com

:80/support/kb/articles/q231/6/19.asp&NoWebContent=1 )
quote:

2003[QUOTE]
no[QUOTE]
back[QUOTE]
benefit[QUOTE]
>
|||Greg,
It sure sounds like a hardware issue, but I don't have any great
suggestions. If you haven't already looked, see if these Knowledge Base
articles help at all:
http://support.microsoft.com:80/sup...s/q231/6/19.asp
(and the articles it references at the bottom)
http://support.microsoft.com/defaul...kb;en-us;268481
It is sounding more like it's server C and not the service pack, but
you never know... It does seem a little strange that restore fails so
easily, yet you have to work to get SQLIOstress to error out.
SK
Greg wrote:
quote:

>Thanks for the reply, Steve.
>I understand that applying the service packs would make the most sense as a
>starting point...but the powers that be do not want me to do that since
>"that's only one that's working".
>The reason I didn't think it was server A is that all interaction between
>Server A and Server B (backup/restore, attach/detach) work fine. Any
>interaction between Server B and Server C result in the same consistency
>problems.
>As an update the SQLIOstress failed out with a series of these errors:
>
>
>605885 and found 601789 in sector 15. Probably a torn page.
>
>Followed by this:
>
>
>--
>
>encountered
>
>--
>I'm guessing that we have a hardware issue. Any thoughts? It seems like i
t
>is up to me to "prove" that the hardware is having issues but I'm not reall
y
>sure what those errors indicate.
>Once again, any help is appreciated.
>-Greg
>
>
>"Steve Kass" <skass@.drew.edu> wrote in message
>news:%23bYZld26DHA.2556@.TK2MSFTNGP09.phx.gbl...
>
>a
>
>Server
>
>then
>
>DBCC
>
>finally
>
>I
>
>to
>
>mentioned,
>
>this
>
>:80/support/kb/articles/q231/6/19.asp&NoWebContent=1 )
>
>2003
>
>no
>
>back
>
>benefit
>
>
>

Can anyone get me out of SSPI / Kerberos / NTLM hell ??

This is to do with SQL Server 2005, but I think it is a more general issue.
We have two SQL Servers, but starting with bog-standard Domain accounts and
both running in Windows Authentication mode. Either account it trust for
delegation, nor are the machines, as we do not need delegation and don't
which to turn it on.
When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
connection and we can see that is it using NTLM for authentication.
When trying to attach to SQL Server #2, using TCP/IP, get an error, "Cannot
generate SSPI context". When trying to attach using Named Pipes, get a
connection and again we can see that it is using NTLM for authentication.
SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
tell) IP Security configured.
We are using Active Directory 2003, running in Native Mode.
What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL Server
#2.
All ideas most welcome
Are the two machines in the same OU ?
There are policy settings that can impact whether NTLM, v1, v2,
are available, and also policies that control the digital communications
signing requirements. If the two are in different OU, or if there are
any GPOs that have their application security group filtered, or if
these policies are left to the local policies to set and they are there
set differently, any one of those could explain your situation.
Assuming that you have looked at the SQL network server and
client library settings and these are in agreement, and that as far as
you can tell the two SQL Server services are otherwise configured
in same/similar way, then it could well be different policy settings.
You could try using resultant set of policy to see the effective
settings on the two servers and compare, particularly the policy
settings in the innermost Security section of Computer policy that
impact NTLM level and signing, and impact network server and
network client behaviors.
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome
|||The machines are both in the same OU. So I'm guessing that the same GPO is
being applied to both machines.
I have just found the "RSoS" snap-in. Had no idea the thing existed.
Could you let me have the names of the policies that effect Kerberos and
NTLM ?
Thanks
"Roger Abell [MVP]" wrote:

> Are the two machines in the same OU ?
> There are policy settings that can impact whether NTLM, v1, v2,
> are available, and also policies that control the digital communications
> signing requirements. If the two are in different OU, or if there are
> any GPOs that have their application security group filtered, or if
> these policies are left to the local policies to set and they are there
> set differently, any one of those could explain your situation.
> Assuming that you have looked at the SQL network server and
> client library settings and these are in agreement, and that as far as
> you can tell the two SQL Server services are otherwise configured
> in same/similar way, then it could well be different policy settings.
> You could try using resultant set of policy to see the effective
> settings on the two servers and compare, particularly the policy
> settings in the innermost Security section of Computer policy that
> impact NTLM level and signing, and impact network server and
> network client behaviors.
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
>
>
|||I suspect Server #2 has an SPN registered for the SQL Server, but the
account running the server has no priviledges over that SPN. Check what SPNs
are registered for the two machines. Use a tool like setspn.exe (available
for download from microsoft download center).
This article is very usefull on troubleshooting these kind of issues:
http://www.microsoft.com/technet/pro.../tkerberr.mspx
This posting is provided "AS IS" with no warranties, and confers no rights.
HTH,
~ Remus Rusanu
SQL Service Broker
http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome
|||Thanks Remus, you were spot on with that.
We hadn't noticed that when we bounced the SQL instance that it was logging
that it couldn't unregister the SPN.
Not 100% sure how we got into that state, but not important.
Thanks again,
Al
"Remus Rusanu [MSFT]" wrote:

> I suspect Server #2 has an SPN registered for the SQL Server, but the
> account running the server has no priviledges over that SPN. Check what SPNs
> are registered for the two machines. Use a tool like setspn.exe (available
> for download from microsoft download center).
> This article is very usefull on troubleshooting these kind of issues:
> http://www.microsoft.com/technet/pro.../tkerberr.mspx
> --
> This posting is provided "AS IS" with no warranties, and confers no rights.
> HTH,
> ~ Remus Rusanu
> SQL Service Broker
> http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
>
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
>
>
|||This one has me in hell too and SPN is my issue as well, but I'm trying to
find a less manual solution because this one keeps popping up.
We're running SQL Server 2005 cluster (active/active). SQL server services
on both nodes is running under a domain account that has been added to the
Domain Admins group. Sporadically, whenever we fail over or reboot the
cluster, the SPN for one of the cluster nodes is lost killing all my apps
that rely on Kerberos authentication. Manually adding an SPN for the SQL
server service account resolves the issue, but this always incurs down time.
Why is the SPN not being created automatically? Permission issue? I've read
that this is only a problem when running the SQL service under a domain
account. Shouldn't being part of the Domain Admin group supply sufficient
permissions to enable the domain account to create the SPN on service startup?
Any help with this is GREATLY appreciated.
James
"Al" wrote:
[vbcol=seagreen]
> Thanks Remus, you were spot on with that.
> We hadn't noticed that when we bounced the SQL instance that it was logging
> that it couldn't unregister the SPN.
> Not 100% sure how we got into that state, but not important.
> Thanks again,
> Al
> "Remus Rusanu [MSFT]" wrote:
|||You would think that it should do the trick, having the account belonging to
the Domain Admin Group. I'm assuming that the SQLserver/service only tries to
register itself when it's running as one of the local system accounts.
When running a failover cluster, I don't think you should be registering the
SPN for both nodes individually. I think you should be registering only the
cluster instance name. Or have I misunderstood what you're been doing to
resolve the problem.
Al
"BeezelDub" wrote:
[vbcol=seagreen]
> This one has me in hell too and SPN is my issue as well, but I'm trying to
> find a less manual solution because this one keeps popping up.
> We're running SQL Server 2005 cluster (active/active). SQL server services
> on both nodes is running under a domain account that has been added to the
> Domain Admins group. Sporadically, whenever we fail over or reboot the
> cluster, the SPN for one of the cluster nodes is lost killing all my apps
> that rely on Kerberos authentication. Manually adding an SPN for the SQL
> server service account resolves the issue, but this always incurs down time.
> Why is the SPN not being created automatically? Permission issue? I've read
> that this is only a problem when running the SQL service under a domain
> account. Shouldn't being part of the Domain Admin group supply sufficient
> permissions to enable the domain account to create the SPN on service startup?
> Any help with this is GREATLY appreciated.
> James
>
>
> "Al" wrote:
|||When Kerberos is failing, a setspn -L for the SQL service domain account
lists...
MSSQLSvc/cluster1.domain.com:1433
MSSQLSvc/cluster1.domain.com
MSSQLSvc/cluster2.domain.com
It's missing the SPN for the host node of the databases that are all of a
sudden unaccessible via Kerberos.
Manually setting the SPN...
setspn -A MSSQLSvc/cluster2.domain.com:1433 sqladmin
Fixes the issue immediately.
To my knowledge, cluster1 has not had this SPN problem but then again, there
are no apps requiring Kerberos hitting databases on that node. In handling
this issue in the past, only cluster2 loses the SPN for whatever reason.
It's nice knowing the fix, but its not THE FIX. Certainly there is some
magic checkbox somewhere that will allow this SPN to be set automatically or
not be lost in the first place.
Thanks for working with me on this!
"Al" wrote:
[vbcol=seagreen]
> You would think that it should do the trick, having the account belonging to
> the Domain Admin Group. I'm assuming that the SQLserver/service only tries to
> register itself when it's running as one of the local system accounts.
> When running a failover cluster, I don't think you should be registering the
> SPN for both nodes individually. I think you should be registering only the
> cluster instance name. Or have I misunderstood what you're been doing to
> resolve the problem.
>
> Al
> "BeezelDub" wrote:
|||I was able to implement this by doing the following within AD (my
service account is not a domain admin):
One time only, for the service account:
1. Service account trusted for delegation (check)
2. Service account is sensitive and cannot be delegated (un-check)
3. Service account - two privileges required, read and write service
principal name, to allow SQL Service startup account to register spn
automatically
Once per each server:
4. Each SQL Server: computer account trusted for delegation
As your account is a domain admin, perhaps step 4 is all that you're
missing.
See MSKB 319723 for more info.
http://support.microsoft.com/kb/319723
Glenn
On Jan 11, 11:46 am, BeezelDub <jam...@.noemail.noemail> wrote:[vbcol=seagreen]
> When Kerberos is failing, a setspn -L for the SQL service domain account
> lists...
> MSSQLSvc/cluster1.domain.com:1433
> MSSQLSvc/cluster1.domain.com
> MSSQLSvc/cluster2.domain.com
> It's missing the SPN for the host node of the databases that are all of a
> sudden unaccessible via Kerberos.
> Manually setting the SPN...
> setspn -A MSSQLSvc/cluster2.domain.com:1433 sqladmin
> Fixes the issue immediately.
> To my knowledge, cluster1 has not had this SPN problem but then again, there
> are no apps requiring Kerberos hitting databases on that node. In handling
> this issue in the past, only cluster2 loses the SPN for whatever reason.
> It's nice knowing the fix, but its not THE FIX. Certainly there is some
> magic checkbox somewhere that will allow this SPN to be set automatically or
> not be lost in the first place.
> Thanks for working with me on this!
> "Al" wrote:
>
>
>
>
>
>
>
>
>
>
>
>
>

Can anyone get me out of SSPI / Kerberos / NTLM hell ??

This is to do with SQL Server 2005, but I think it is a more general issue.
We have two SQL Servers, but starting with bog-standard Domain accounts and
both running in Windows Authentication mode. Either account it trust for
delegation, nor are the machines, as we do not need delegation and don't
which to turn it on.
When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
connection and we can see that is it using NTLM for authentication.
When trying to attach to SQL Server #2, using TCP/IP, get an error, "Cannot
generate SSPI context". When trying to attach using Named Pipes, get a
connection and again we can see that it is using NTLM for authentication.
SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
tell) IP Security configured.
We are using Active Directory 2003, running in Native Mode.
What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL Server
#2.
All ideas most welcome
Are the two machines in the same OU ?
There are policy settings that can impact whether NTLM, v1, v2,
are available, and also policies that control the digital communications
signing requirements. If the two are in different OU, or if there are
any GPOs that have their application security group filtered, or if
these policies are left to the local policies to set and they are there
set differently, any one of those could explain your situation.
Assuming that you have looked at the SQL network server and
client library settings and these are in agreement, and that as far as
you can tell the two SQL Server services are otherwise configured
in same/similar way, then it could well be different policy settings.
You could try using resultant set of policy to see the effective
settings on the two servers and compare, particularly the policy
settings in the innermost Security section of Computer policy that
impact NTLM level and signing, and impact network server and
network client behaviors.
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome
|||The machines are both in the same OU. So I'm guessing that the same GPO is
being applied to both machines.
I have just found the "RSoS" snap-in. Had no idea the thing existed.
Could you let me have the names of the policies that effect Kerberos and
NTLM ?
Thanks
"Roger Abell [MVP]" wrote:

> Are the two machines in the same OU ?
> There are policy settings that can impact whether NTLM, v1, v2,
> are available, and also policies that control the digital communications
> signing requirements. If the two are in different OU, or if there are
> any GPOs that have their application security group filtered, or if
> these policies are left to the local policies to set and they are there
> set differently, any one of those could explain your situation.
> Assuming that you have looked at the SQL network server and
> client library settings and these are in agreement, and that as far as
> you can tell the two SQL Server services are otherwise configured
> in same/similar way, then it could well be different policy settings.
> You could try using resultant set of policy to see the effective
> settings on the two servers and compare, particularly the policy
> settings in the innermost Security section of Computer policy that
> impact NTLM level and signing, and impact network server and
> network client behaviors.
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
>
>
|||I suspect Server #2 has an SPN registered for the SQL Server, but the
account running the server has no priviledges over that SPN. Check what SPNs
are registered for the two machines. Use a tool like setspn.exe (available
for download from microsoft download center).
This article is very usefull on troubleshooting these kind of issues:
http://www.microsoft.com/technet/pro.../tkerberr.mspx
This posting is provided "AS IS" with no warranties, and confers no rights.
HTH,
~ Remus Rusanu
SQL Service Broker
http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome
|||Thanks Remus, you were spot on with that.
We hadn't noticed that when we bounced the SQL instance that it was logging
that it couldn't unregister the SPN.
Not 100% sure how we got into that state, but not important.
Thanks again,
Al
"Remus Rusanu [MSFT]" wrote:

> I suspect Server #2 has an SPN registered for the SQL Server, but the
> account running the server has no priviledges over that SPN. Check what SPNs
> are registered for the two machines. Use a tool like setspn.exe (available
> for download from microsoft download center).
> This article is very usefull on troubleshooting these kind of issues:
> http://www.microsoft.com/technet/pro.../tkerberr.mspx
> --
> This posting is provided "AS IS" with no warranties, and confers no rights.
> HTH,
> ~ Remus Rusanu
> SQL Service Broker
> http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
>
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
>
>
|||This one has me in hell too and SPN is my issue as well, but I'm trying to
find a less manual solution because this one keeps popping up.
We're running SQL Server 2005 cluster (active/active). SQL server services
on both nodes is running under a domain account that has been added to the
Domain Admins group. Sporadically, whenever we fail over or reboot the
cluster, the SPN for one of the cluster nodes is lost killing all my apps
that rely on Kerberos authentication. Manually adding an SPN for the SQL
server service account resolves the issue, but this always incurs down time.
Why is the SPN not being created automatically? Permission issue? I've read
that this is only a problem when running the SQL service under a domain
account. Shouldn't being part of the Domain Admin group supply sufficient
permissions to enable the domain account to create the SPN on service startup?
Any help with this is GREATLY appreciated.
James
"Al" wrote:
[vbcol=seagreen]
> Thanks Remus, you were spot on with that.
> We hadn't noticed that when we bounced the SQL instance that it was logging
> that it couldn't unregister the SPN.
> Not 100% sure how we got into that state, but not important.
> Thanks again,
> Al
> "Remus Rusanu [MSFT]" wrote:
|||You would think that it should do the trick, having the account belonging to
the Domain Admin Group. I'm assuming that the SQLserver/service only tries to
register itself when it's running as one of the local system accounts.
When running a failover cluster, I don't think you should be registering the
SPN for both nodes individually. I think you should be registering only the
cluster instance name. Or have I misunderstood what you're been doing to
resolve the problem.
Al
"BeezelDub" wrote:
[vbcol=seagreen]
> This one has me in hell too and SPN is my issue as well, but I'm trying to
> find a less manual solution because this one keeps popping up.
> We're running SQL Server 2005 cluster (active/active). SQL server services
> on both nodes is running under a domain account that has been added to the
> Domain Admins group. Sporadically, whenever we fail over or reboot the
> cluster, the SPN for one of the cluster nodes is lost killing all my apps
> that rely on Kerberos authentication. Manually adding an SPN for the SQL
> server service account resolves the issue, but this always incurs down time.
> Why is the SPN not being created automatically? Permission issue? I've read
> that this is only a problem when running the SQL service under a domain
> account. Shouldn't being part of the Domain Admin group supply sufficient
> permissions to enable the domain account to create the SPN on service startup?
> Any help with this is GREATLY appreciated.
> James
>
>
> "Al" wrote:
|||When Kerberos is failing, a setspn -L for the SQL service domain account
lists...
MSSQLSvc/cluster1.domain.com:1433
MSSQLSvc/cluster1.domain.com
MSSQLSvc/cluster2.domain.com
It's missing the SPN for the host node of the databases that are all of a
sudden unaccessible via Kerberos.
Manually setting the SPN...
setspn -A MSSQLSvc/cluster2.domain.com:1433 sqladmin
Fixes the issue immediately.
To my knowledge, cluster1 has not had this SPN problem but then again, there
are no apps requiring Kerberos hitting databases on that node. In handling
this issue in the past, only cluster2 loses the SPN for whatever reason.
It's nice knowing the fix, but its not THE FIX. Certainly there is some
magic checkbox somewhere that will allow this SPN to be set automatically or
not be lost in the first place.
Thanks for working with me on this!
"Al" wrote:
[vbcol=seagreen]
> You would think that it should do the trick, having the account belonging to
> the Domain Admin Group. I'm assuming that the SQLserver/service only tries to
> register itself when it's running as one of the local system accounts.
> When running a failover cluster, I don't think you should be registering the
> SPN for both nodes individually. I think you should be registering only the
> cluster instance name. Or have I misunderstood what you're been doing to
> resolve the problem.
>
> Al
> "BeezelDub" wrote:
|||I was able to implement this by doing the following within AD (my
service account is not a domain admin):
One time only, for the service account:
1. Service account trusted for delegation (check)
2. Service account is sensitive and cannot be delegated (un-check)
3. Service account - two privileges required, read and write service
principal name, to allow SQL Service startup account to register spn
automatically
Once per each server:
4. Each SQL Server: computer account trusted for delegation
As your account is a domain admin, perhaps step 4 is all that you're
missing.
See MSKB 319723 for more info.
http://support.microsoft.com/kb/319723
Glenn
On Jan 11, 11:46 am, BeezelDub <jam...@.noemail.noemail> wrote:[vbcol=seagreen]
> When Kerberos is failing, a setspn -L for the SQL service domain account
> lists...
> MSSQLSvc/cluster1.domain.com:1433
> MSSQLSvc/cluster1.domain.com
> MSSQLSvc/cluster2.domain.com
> It's missing the SPN for the host node of the databases that are all of a
> sudden unaccessible via Kerberos.
> Manually setting the SPN...
> setspn -A MSSQLSvc/cluster2.domain.com:1433 sqladmin
> Fixes the issue immediately.
> To my knowledge, cluster1 has not had this SPN problem but then again, there
> are no apps requiring Kerberos hitting databases on that node. In handling
> this issue in the past, only cluster2 loses the SPN for whatever reason.
> It's nice knowing the fix, but its not THE FIX. Certainly there is some
> magic checkbox somewhere that will allow this SPN to be set automatically or
> not be lost in the first place.
> Thanks for working with me on this!
> "Al" wrote:
>
>
>
>
>
>
>
>
>
>
>
>
>

Can anyone get me out of SSPI / Kerberos / NTLM hell ??

This is to do with SQL Server 2005, but I think it is a more general issue.
We have two SQL Servers, but starting with bog-standard Domain accounts and
both running in Windows Authentication mode. Either account it trust for
delegation, nor are the machines, as we do not need delegation and don't
which to turn it on.
When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
connection and we can see that is it using NTLM for authentication.
When trying to attach to SQL Server #2, using TCP/IP, get an error, "Cannot
generate SSPI context". When trying to attach using Named Pipes, get a
connection and again we can see that it is using NTLM for authentication.
SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
tell) IP Security configured.
We are using Active Directory 2003, running in Native Mode.
What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL Server
#2.
All ideas most welcomeAre the two machines in the same OU ?
There are policy settings that can impact whether NTLM, v1, v2,
are available, and also policies that control the digital communications
signing requirements. If the two are in different OU, or if there are
any GPOs that have their application security group filtered, or if
these policies are left to the local policies to set and they are there
set differently, any one of those could explain your situation.
Assuming that you have looked at the SQL network server and
client library settings and these are in agreement, and that as far as
you can tell the two SQL Server services are otherwise configured
in same/similar way, then it could well be different policy settings.
You could try using resultant set of policy to see the effective
settings on the two servers and compare, particularly the policy
settings in the innermost Security section of Computer policy that
impact NTLM level and signing, and impact network server and
network client behaviors.
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome|||The machines are both in the same OU. So I'm guessing that the same GPO is
being applied to both machines.
I have just found the "RSoS" snap-in. Had no idea the thing existed.
Could you let me have the names of the policies that effect Kerberos and
NTLM ?
Thanks
"Roger Abell [MVP]" wrote:
> Are the two machines in the same OU ?
> There are policy settings that can impact whether NTLM, v1, v2,
> are available, and also policies that control the digital communications
> signing requirements. If the two are in different OU, or if there are
> any GPOs that have their application security group filtered, or if
> these policies are left to the local policies to set and they are there
> set differently, any one of those could explain your situation.
> Assuming that you have looked at the SQL network server and
> client library settings and these are in agreement, and that as far as
> you can tell the two SQL Server services are otherwise configured
> in same/similar way, then it could well be different policy settings.
> You could try using resultant set of policy to see the effective
> settings on the two servers and compare, particularly the policy
> settings in the innermost Security section of Computer policy that
> impact NTLM level and signing, and impact network server and
> network client behaviors.
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> > This is to do with SQL Server 2005, but I think it is a more general
> > issue.
> >
> > We have two SQL Servers, but starting with bog-standard Domain accounts
> > and
> > both running in Windows Authentication mode. Either account it trust for
> > delegation, nor are the machines, as we do not need delegation and don't
> > which to turn it on.
> >
> > When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> > connection and we can see that is it using NTLM for authentication.
> >
> > When trying to attach to SQL Server #2, using TCP/IP, get an error,
> > "Cannot
> > generate SSPI context". When trying to attach using Named Pipes, get a
> > connection and again we can see that it is using NTLM for authentication.
> >
> > SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> > tell) IP Security configured.
> >
> > We are using Active Directory 2003, running in Native Mode.
> >
> > What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> > Server
> > #2.
> >
> > All ideas most welcome
>
>|||I suspect Server #2 has an SPN registered for the SQL Server, but the
account running the server has no priviledges over that SPN. Check what SPNs
are registered for the two machines. Use a tool like setspn.exe (available
for download from microsoft download center).
This article is very usefull on troubleshooting these kind of issues:
http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/security/tkerberr.mspx
--
This posting is provided "AS IS" with no warranties, and confers no rights.
HTH,
~ Remus Rusanu
SQL Service Broker
http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome|||Thanks Remus, you were spot on with that.
We hadn't noticed that when we bounced the SQL instance that it was logging
that it couldn't unregister the SPN.
Not 100% sure how we got into that state, but not important.
Thanks again,
Al
"Remus Rusanu [MSFT]" wrote:
> I suspect Server #2 has an SPN registered for the SQL Server, but the
> account running the server has no priviledges over that SPN. Check what SPNs
> are registered for the two machines. Use a tool like setspn.exe (available
> for download from microsoft download center).
> This article is very usefull on troubleshooting these kind of issues:
> http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/security/tkerberr.mspx
> --
> This posting is provided "AS IS" with no warranties, and confers no rights.
> HTH,
> ~ Remus Rusanu
> SQL Service Broker
> http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
>
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> > This is to do with SQL Server 2005, but I think it is a more general
> > issue.
> >
> > We have two SQL Servers, but starting with bog-standard Domain accounts
> > and
> > both running in Windows Authentication mode. Either account it trust for
> > delegation, nor are the machines, as we do not need delegation and don't
> > which to turn it on.
> >
> > When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> > connection and we can see that is it using NTLM for authentication.
> >
> > When trying to attach to SQL Server #2, using TCP/IP, get an error,
> > "Cannot
> > generate SSPI context". When trying to attach using Named Pipes, get a
> > connection and again we can see that it is using NTLM for authentication.
> >
> > SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> > tell) IP Security configured.
> >
> > We are using Active Directory 2003, running in Native Mode.
> >
> > What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> > Server
> > #2.
> >
> > All ideas most welcome
>
>|||This one has me in hell too and SPN is my issue as well, but I'm trying to
find a less manual solution because this one keeps popping up.
We're running SQL Server 2005 cluster (active/active). SQL server services
on both nodes is running under a domain account that has been added to the
Domain Admins group. Sporadically, whenever we fail over or reboot the
cluster, the SPN for one of the cluster nodes is lost killing all my apps
that rely on Kerberos authentication. Manually adding an SPN for the SQL
server service account resolves the issue, but this always incurs down time.
Why is the SPN not being created automatically? Permission issue? I've read
that this is only a problem when running the SQL service under a domain
account. Shouldn't being part of the Domain Admin group supply sufficient
permissions to enable the domain account to create the SPN on service startup?
Any help with this is GREATLY appreciated.
James
"Al" wrote:
> Thanks Remus, you were spot on with that.
> We hadn't noticed that when we bounced the SQL instance that it was logging
> that it couldn't unregister the SPN.
> Not 100% sure how we got into that state, but not important.
> Thanks again,
> Al
> "Remus Rusanu [MSFT]" wrote:
> > I suspect Server #2 has an SPN registered for the SQL Server, but the
> > account running the server has no priviledges over that SPN. Check what SPNs
> > are registered for the two machines. Use a tool like setspn.exe (available
> > for download from microsoft download center).
> >
> > This article is very usefull on troubleshooting these kind of issues:
> > http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/security/tkerberr.mspx
> >
> > --
> > This posting is provided "AS IS" with no warranties, and confers no rights.
> >
> > HTH,
> > ~ Remus Rusanu
> >
> > SQL Service Broker
> > http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
> >
> >
> > "Al" <Al@.discussions.microsoft.com> wrote in message
> > news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> > > This is to do with SQL Server 2005, but I think it is a more general
> > > issue.
> > >
> > > We have two SQL Servers, but starting with bog-standard Domain accounts
> > > and
> > > both running in Windows Authentication mode. Either account it trust for
> > > delegation, nor are the machines, as we do not need delegation and don't
> > > which to turn it on.
> > >
> > > When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> > > connection and we can see that is it using NTLM for authentication.
> > >
> > > When trying to attach to SQL Server #2, using TCP/IP, get an error,
> > > "Cannot
> > > generate SSPI context". When trying to attach using Named Pipes, get a
> > > connection and again we can see that it is using NTLM for authentication.
> > >
> > > SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> > > tell) IP Security configured.
> > >
> > > We are using Active Directory 2003, running in Native Mode.
> > >
> > > What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> > > Server
> > > #2.
> > >
> > > All ideas most welcome
> >
> >
> >|||You would think that it should do the trick, having the account belonging to
the Domain Admin Group. I'm assuming that the SQLserver/service only tries to
register itself when it's running as one of the local system accounts.
When running a failover cluster, I don't think you should be registering the
SPN for both nodes individually. I think you should be registering only the
cluster instance name. Or have I misunderstood what you're been doing to
resolve the problem.
Al
"BeezelDub" wrote:
> This one has me in hell too and SPN is my issue as well, but I'm trying to
> find a less manual solution because this one keeps popping up.
> We're running SQL Server 2005 cluster (active/active). SQL server services
> on both nodes is running under a domain account that has been added to the
> Domain Admins group. Sporadically, whenever we fail over or reboot the
> cluster, the SPN for one of the cluster nodes is lost killing all my apps
> that rely on Kerberos authentication. Manually adding an SPN for the SQL
> server service account resolves the issue, but this always incurs down time.
> Why is the SPN not being created automatically? Permission issue? I've read
> that this is only a problem when running the SQL service under a domain
> account. Shouldn't being part of the Domain Admin group supply sufficient
> permissions to enable the domain account to create the SPN on service startup?
> Any help with this is GREATLY appreciated.
> James
>
>
> "Al" wrote:
> > Thanks Remus, you were spot on with that.
> > We hadn't noticed that when we bounced the SQL instance that it was logging
> > that it couldn't unregister the SPN.
> >
> > Not 100% sure how we got into that state, but not important.
> >
> > Thanks again,
> >
> > Al
> >
> > "Remus Rusanu [MSFT]" wrote:
> >
> > > I suspect Server #2 has an SPN registered for the SQL Server, but the
> > > account running the server has no priviledges over that SPN. Check what SPNs
> > > are registered for the two machines. Use a tool like setspn.exe (available
> > > for download from microsoft download center).
> > >
> > > This article is very usefull on troubleshooting these kind of issues:
> > > http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/security/tkerberr.mspx
> > >
> > > --
> > > This posting is provided "AS IS" with no warranties, and confers no rights.
> > >
> > > HTH,
> > > ~ Remus Rusanu
> > >
> > > SQL Service Broker
> > > http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
> > >
> > >
> > > "Al" <Al@.discussions.microsoft.com> wrote in message
> > > news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> > > > This is to do with SQL Server 2005, but I think it is a more general
> > > > issue.
> > > >
> > > > We have two SQL Servers, but starting with bog-standard Domain accounts
> > > > and
> > > > both running in Windows Authentication mode. Either account it trust for
> > > > delegation, nor are the machines, as we do not need delegation and don't
> > > > which to turn it on.
> > > >
> > > > When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> > > > connection and we can see that is it using NTLM for authentication.
> > > >
> > > > When trying to attach to SQL Server #2, using TCP/IP, get an error,
> > > > "Cannot
> > > > generate SSPI context". When trying to attach using Named Pipes, get a
> > > > connection and again we can see that it is using NTLM for authentication.
> > > >
> > > > SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> > > > tell) IP Security configured.
> > > >
> > > > We are using Active Directory 2003, running in Native Mode.
> > > >
> > > > What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> > > > Server
> > > > #2.
> > > >
> > > > All ideas most welcome
> > >
> > >
> > >|||When Kerberos is failing, a setspn -L for the SQL service domain account
lists...
MSSQLSvc/cluster1.domain.com:1433
MSSQLSvc/cluster1.domain.com
MSSQLSvc/cluster2.domain.com
It's missing the SPN for the host node of the databases that are all of a
sudden unaccessible via Kerberos.
Manually setting the SPN...
setspn -A MSSQLSvc/cluster2.domain.com:1433 sqladmin
Fixes the issue immediately.
To my knowledge, cluster1 has not had this SPN problem but then again, there
are no apps requiring Kerberos hitting databases on that node. In handling
this issue in the past, only cluster2 loses the SPN for whatever reason.
It's nice knowing the fix, but its not THE FIX. Certainly there is some
magic checkbox somewhere that will allow this SPN to be set automatically or
not be lost in the first place.
Thanks for working with me on this!
"Al" wrote:
> You would think that it should do the trick, having the account belonging to
> the Domain Admin Group. I'm assuming that the SQLserver/service only tries to
> register itself when it's running as one of the local system accounts.
> When running a failover cluster, I don't think you should be registering the
> SPN for both nodes individually. I think you should be registering only the
> cluster instance name. Or have I misunderstood what you're been doing to
> resolve the problem.
>
> Al
> "BeezelDub" wrote:
> > This one has me in hell too and SPN is my issue as well, but I'm trying to
> > find a less manual solution because this one keeps popping up.
> >
> > We're running SQL Server 2005 cluster (active/active). SQL server services
> > on both nodes is running under a domain account that has been added to the
> > Domain Admins group. Sporadically, whenever we fail over or reboot the
> > cluster, the SPN for one of the cluster nodes is lost killing all my apps
> > that rely on Kerberos authentication. Manually adding an SPN for the SQL
> > server service account resolves the issue, but this always incurs down time.
> >
> > Why is the SPN not being created automatically? Permission issue? I've read
> > that this is only a problem when running the SQL service under a domain
> > account. Shouldn't being part of the Domain Admin group supply sufficient
> > permissions to enable the domain account to create the SPN on service startup?
> >
> > Any help with this is GREATLY appreciated.
> >
> > James
> >
> >
> >
> >
> > "Al" wrote:
> >
> > > Thanks Remus, you were spot on with that.
> > > We hadn't noticed that when we bounced the SQL instance that it was logging
> > > that it couldn't unregister the SPN.
> > >
> > > Not 100% sure how we got into that state, but not important.
> > >
> > > Thanks again,
> > >
> > > Al
> > >
> > > "Remus Rusanu [MSFT]" wrote:
> > >
> > > > I suspect Server #2 has an SPN registered for the SQL Server, but the
> > > > account running the server has no priviledges over that SPN. Check what SPNs
> > > > are registered for the two machines. Use a tool like setspn.exe (available
> > > > for download from microsoft download center).
> > > >
> > > > This article is very usefull on troubleshooting these kind of issues:
> > > > http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/security/tkerberr.mspx
> > > >
> > > > --
> > > > This posting is provided "AS IS" with no warranties, and confers no rights.
> > > >
> > > > HTH,
> > > > ~ Remus Rusanu
> > > >
> > > > SQL Service Broker
> > > > http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
> > > >
> > > >
> > > > "Al" <Al@.discussions.microsoft.com> wrote in message
> > > > news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> > > > > This is to do with SQL Server 2005, but I think it is a more general
> > > > > issue.
> > > > >
> > > > > We have two SQL Servers, but starting with bog-standard Domain accounts
> > > > > and
> > > > > both running in Windows Authentication mode. Either account it trust for
> > > > > delegation, nor are the machines, as we do not need delegation and don't
> > > > > which to turn it on.
> > > > >
> > > > > When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> > > > > connection and we can see that is it using NTLM for authentication.
> > > > >
> > > > > When trying to attach to SQL Server #2, using TCP/IP, get an error,
> > > > > "Cannot
> > > > > generate SSPI context". When trying to attach using Named Pipes, get a
> > > > > connection and again we can see that it is using NTLM for authentication.
> > > > >
> > > > > SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> > > > > tell) IP Security configured.
> > > > >
> > > > > We are using Active Directory 2003, running in Native Mode.
> > > > >
> > > > > What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> > > > > Server
> > > > > #2.
> > > > >
> > > > > All ideas most welcome
> > > >
> > > >
> > > >

Can anyone get me out of SSPI / Kerberos / NTLM hell ??

This is to do with SQL Server 2005, but I think it is a more general issue.
We have two SQL Servers, but starting with bog-standard Domain accounts and
both running in Windows Authentication mode. Either account it trust for
delegation, nor are the machines, as we do not need delegation and don't
which to turn it on.
When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
connection and we can see that is it using NTLM for authentication.
When trying to attach to SQL Server #2, using TCP/IP, get an error, "Cannot
generate SSPI context". When trying to attach using Named Pipes, get a
connection and again we can see that it is using NTLM for authentication.
SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
tell) IP Security configured.
We are using Active Directory 2003, running in Native Mode.
What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL Server
#2.
All ideas most welcomeAre the two machines in the same OU ?
There are policy settings that can impact whether NTLM, v1, v2,
are available, and also policies that control the digital communications
signing requirements. If the two are in different OU, or if there are
any GPOs that have their application security group filtered, or if
these policies are left to the local policies to set and they are there
set differently, any one of those could explain your situation.
Assuming that you have looked at the SQL network server and
client library settings and these are in agreement, and that as far as
you can tell the two SQL Server services are otherwise configured
in same/similar way, then it could well be different policy settings.
You could try using resultant set of policy to see the effective
settings on the two servers and compare, particularly the policy
settings in the innermost Security section of Computer policy that
impact NTLM level and signing, and impact network server and
network client behaviors.
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome|||The machines are both in the same OU. So I'm guessing that the same GPO is
being applied to both machines.
I have just found the "RSoS" snap-in. Had no idea the thing existed.
Could you let me have the names of the policies that effect Kerberos and
NTLM ?
Thanks
"Roger Abell [MVP]" wrote:

> Are the two machines in the same OU ?
> There are policy settings that can impact whether NTLM, v1, v2,
> are available, and also policies that control the digital communications
> signing requirements. If the two are in different OU, or if there are
> any GPOs that have their application security group filtered, or if
> these policies are left to the local policies to set and they are there
> set differently, any one of those could explain your situation.
> Assuming that you have looked at the SQL network server and
> client library settings and these are in agreement, and that as far as
> you can tell the two SQL Server services are otherwise configured
> in same/similar way, then it could well be different policy settings.
> You could try using resultant set of policy to see the effective
> settings on the two servers and compare, particularly the policy
> settings in the innermost Security section of Computer policy that
> impact NTLM level and signing, and impact network server and
> network client behaviors.
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
>
>|||I suspect Server #2 has an SPN registered for the SQL Server, but the
account running the server has no priviledges over that SPN. Check what SPNs
are registered for the two machines. Use a tool like setspn.exe (available
for download from microsoft download center).
This article is very usefull on troubleshooting these kind of issues:
http://www.microsoft.com/technet/pr...y/tkerberr.mspx
This posting is provided "AS IS" with no warranties, and confers no rights.
HTH,
~ Remus Rusanu
SQL Service Broker
http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
"Al" <Al@.discussions.microsoft.com> wrote in message
news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
> This is to do with SQL Server 2005, but I think it is a more general
> issue.
> We have two SQL Servers, but starting with bog-standard Domain accounts
> and
> both running in Windows Authentication mode. Either account it trust for
> delegation, nor are the machines, as we do not need delegation and don't
> which to turn it on.
> When trying to attach to SQL Server #1, with Windows using TCP/IP, get a
> connection and we can see that is it using NTLM for authentication.
> When trying to attach to SQL Server #2, using TCP/IP, get an error,
> "Cannot
> generate SSPI context". When trying to attach using Named Pipes, get a
> connection and again we can see that it is using NTLM for authentication.
> SQL Server #2 does have TCP/IP enabled. It does not have (as far as I can
> tell) IP Security configured.
> We are using Active Directory 2003, running in Native Mode.
> What would allow SQL Server #1 to allow NTLM over TCP/IP, but not SQL
> Server
> #2.
> All ideas most welcome|||Thanks Remus, you were spot on with that.
We hadn't noticed that when we bounced the SQL instance that it was logging
that it couldn't unregister the SPN.
Not 100% sure how we got into that state, but not important.
Thanks again,
Al
"Remus Rusanu [MSFT]" wrote:

> I suspect Server #2 has an SPN registered for the SQL Server, but the
> account running the server has no priviledges over that SPN. Check what SP
Ns
> are registered for the two machines. Use a tool like setspn.exe (available
> for download from microsoft download center).
> This article is very usefull on troubleshooting these kind of issues:
> http://www.microsoft.com/technet/pr...y/tkerberr.mspx
> --
> This posting is provided "AS IS" with no warranties, and confers no rights
.
> HTH,
> ~ Remus Rusanu
> SQL Service Broker
> http://msdn2.microsoft.com/en-us/library/ms166043(en-US,SQL.90).aspx
>
> "Al" <Al@.discussions.microsoft.com> wrote in message
> news:50D9C968-5AB6-4D1C-9D7D-5E8BD538785F@.microsoft.com...
>
>