Re: Unable to Apply SP4 to SQL 2000 Cluster (new Node)

Anthony,

Ugh, sorry for not replying earlier. I'm just getting over whatever crud
has had me down for the week.

The guidance that is in the SP documentation sees to contradict what you are
saying (not that M$ documentation is always correct).

[quote]
If you need to rebuild a node in the failover cluster:

1. Rebuild the node in the failover cluster. For more information about
rebuilding a node, see "How to recover from failover cluster failure in
Scenario 1" in SQL Server 2000 Books Online.

2. Run the original SQL Server 2000 Setup program to add the node back to
the failover cluster.

3. Run Database Components SP4 Setup [bold]on the node you have
added[/bold]. This setup process updates to SP4 only the binaries on the new
node.

[bold]Note If you run Setup from the node where the virtual server is
running, you must reapply SP4 to all the nodes. You must also run the
database upgrade scripts again.[/bold]
[/quote]

This is the process that I went through in my test environment and it worked
fine. It's not working in the production environment which is what is
mystifying me.

A fundamental concern that I have is with downtime; I get a very limited
amount of time to do maintenance each month. Starting to apply the SP from
the active node appears to cause downtime; it stops and restarts the database
service a couple of times. I was hoping to avoid that problem by pursuing
the binaries only install method (again, it worked in the test environment
just fine).

Okay, I think it's time to open a ticket with MS and see where I get with
that route. I will post an update when I do finally get a resolution (but it
may be a couple of weeks because I'm out on vacation next week).

Thanks for your time and effort and patience. I'm really not trying to be
difficult or to ignore your advice. I'm just trying to figure out why the
same steps that worked in the test environment are not working in the
production environment.

Best regards and Happy Holidays,

hmscott




"Anthony Thomas" wrote:

"All cluster disks available to this virtual server are owned by other
nodes."

When do you receive it?

While attempting to install SP4 (or SP3), it comes up immediately after I
enter the name of the virtual instance to be upgraded (I start the service
pack upgrade from the node that is being patched in accordance with the
guidelines in the SP documentation).

Where did you locate the message?

It's a pop-up message on the console. It also appears in the sqlsp.log
file. (? I'm pretty sure that's it, but I'm home sick and the battery on
my
laptop finally gave out).

You have to run a "binary only" patch/SP deployment from the node that
currently owns the resources. You receive this error message when the node
you are on does not own the resources. You just need to install from the
other node; it will detect that you are only patching the others.

I suspect that you will have trouble applying SP4 directly. You can try it,
but in my experience when trying to bring another node up to the same level
as an existing node (as opposed to a new build), you will need to apply SP3a
first. The reason is that the remote unattended installation will fail
because of the pre-SP3 warning is launched in interactive mode. Since it is
a remote installation, there is no interactive users. Moreover, the dialog
is launched because there is a bug prior to 2148 and 2151 that misreads the
build number, truncating one of the digits (SP4 increased the build to a
4-digit number). So, instead of 2039 (SP4 build), it reads only 203 or 039,
both of which is less than the SP3a build of 718.

On Win2K3, security features will disable the TCP/IP communication until
SP3a is successfully deployed. You will need to create client-side aliases
on member nodes for the instance you are working with defined using Named
Pipes. After SP3a is deployed, you can switch these to use TCP/IP and the
port assigned, unless you are using Dynamic Ports, which I strongly
discourage due to the UDP 1434 issues.

You are also having the same issue installing instance 1 an 2 on node 4,
currently only at RTM binaries. The same issue applies as above: you have
to be on one of the nodes that can bring SQL Server online and apply SP3a,
SP4, HF2040, and recommend HF2187.

When you are done, you will also want to export the
HKLM\SOFTWARE\Microsoft\Microsoft SQL Server and MSSQLSERVER registry keys
from one of the original nodes and merge them on nodes 3 and 4. The add
node setup does not copy these and they do not replicate completely through
the cluster quorum.

I can include some KB and MSDN articles, but there is no one top to bottom
recommendation.

Failover Cluster Troubleshooting
http://msdn2.microsoft.com/en-us/library/aa196696(sql.80).aspx

Installation of a named instance of SQL Server 2000 virtual server on a
Windows Server 2003-based cluster fails
http://support.microsoft.com/kb/815431/en-us

FIX: You receive an error message when you run SQL Server Setup to install
SQL Server 2000 Service Pack 4 in binary-only mode on a failover cluster
http://support.microsoft.com/kb/905286/en-us

Sincerely,


Anthony Thomas


--

"hmscott" <hmscott@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:2740AE70-27C8-4FDF-A96A-D8A033FEE351@xxxxxxxxxxxxxxxx
Anthony,

First, many thanks for your patience and assistance.

To clear up the issue of "on board" drives: each server has on on-board
RAID
controller and the specified number of drives configured as I previously
mentioned. These are NOT the shared drives. Each server is configured
with

C:\ (system, boot, paging file),
E:\ (paging file, SQL Binaries and other applications, like Oracle client
drivers)
F:\ Drive (which is used for a paging file)

I mentioned the on board drives because I though they might be
contributing
to the problem. On the IBMs, there is 1 virtual disk with 3 logical
partitions. On the two Dells, there are two virtual disks, one with a
single
partition and one with two partitions.

In addition to the on-board drives, each server has two HBA connections to
our IBM SAN. All shared drives are on the SAN. There are a total of 14
shared drives (in five cluster groups).

Cluster Group
- 1 disk
- IP address &tc

MS DTC Group
- 1 Disk
- IP Address &tc

DBC01 (SQL Instance 1)
- 3 Disks
- IP Address and SQL Resources

DBC02 (SQL Instance 2)
- 3 Disks
- IP Address and SQL Resources

DBC03 (SQL Instance 3)
- 6 disks
- IP Address and SQL Resources

Now on to the other questions:

You then installed a 2nd instance of SQL Server. May I ask which member
nodes were included in the SQL Server Setup? All 3 nodes?

Yes, I was able to go through the entire set up process for adding the
second SQL instance and make in work on all three nodes (including adding
SP4
and Hf 2040).

Can all member nodes take ownership of this second instance?

At the time of installation of the third node/second SQL instance, all
three
nodes were able to "own" the second instance and all three nodes were
patched
to the same version (ie, 2040). Currently the 4th node can "own" the
second
instance, but only at RTM patch level.

Now, you are trying to tackle including the new 3rd node as a potential
owner of the 1st SQL Server instance, but are running into disk
ownership
messages.

What is the exact error message?

"All cluster disks available to this virtual server are owned by other
nodes."

When do you receive it?

While attempting to install SP4 (or SP3), it comes up immediately after I
enter the name of the virtual instance to be upgraded (I start the service
pack upgrade from the node that is being patched in accordance with the
guidelines in the SP documentation).

Where did you locate the message?

It's a pop-up message on the console. It also appears in the sqlsp.log
file. (? I'm pretty sure that's it, but I'm home sick and the battery on
my
laptop finally gave out).

What method did you use to place the RTM bits on the 3rd node?

I started the RTM setup file (while logged in to the new node), specifed
Advanced Options / Maintain Failover Cluster and added the new node to the
list of "configured" servers. This was done while the 1st instance was on
line, but while it was owned by another node.

You will not be able to host disk or services on node 3 for SS instance
one until
the bits are at the same patch level. The cluster service will not
allow you to
move the group if any resource is not a designated owner. If all of
them are,
then you could offline the SQL Server resources, and then try to take
ownership
and only bring the disks, IP, and Network Name online.

I am able to move the SQL Resouces to the new Node (Move Group / Node 3).
However, when it starts, it starts in "protected mode" (there is an entry
in
the SQL Server Error log indicating that SQL has started in "protected
mode".
Also, it is only listening on Named Pipes, not TCP/IP (which I think is
the
definition of Protected Mode?).

Is this when you get your error messages?
No (see note above).

If so, then it is probably a disk signature collision on node 3. If
this is the case,
examine HKLM\SYSTEM\MountedDevices.

I will check this reg entry when I get into the office, but I don't think
this is the source of the problem.

All disclaimers apply: but delete any shared resources, both GUID and
Drive letter
entries. Reboot the server, and then try to take ownership again
without
launching the LDM Administrator (remember, the SQL Server resources will
not
come online until the bits are at the same patch level).

Please clarify LDM Administrator? Local Disk Manager?

If this is not the case, then I will need the answers to the above
questions.

I'm not sure which questions you were referring to. I've tried to answer
all of them. Apologies in advance if I missed any.


Now, when the 4th cluster node was added, were you again successful in
adding it as a cluster node member?

I ran into the same issues as with Node 3 (ie, I had to use the Advance
configuration option on the Cluster Setup Wizard). But other than that,
the
process worked just fine.


When you installed the 3rd SQL Server instance, which cluster nodes did
you add
as potential owners?

I specified all nodes as potential owners. I was able to load the 3rd
instance on all nodes including SP4/2040.


Are you again attempting to add node 4 as a potential owner of SQL
Server
instance 1?
Yes.

And you are receiving the same error messages as on node 3?
Yes. To be exact, I van load the RTM binaries, but when I go to apply SP4
(or SP3), I get the error message: "All cluster disks available to this
virtual server are owned by other nodes."


Is it only the 1st SQL Server instance that you are have trouble with?
No. On Node 4 I receive the same error message when trying to apply SP4
(or
SP3) for SQL instances 2 and 3.


To summarize:

Node 1: Is correctly configured to run SQL 2000 SP4 for all three virtual
instances (SQL 1, 2 and 3)

Node 2: Is correctly configured to run SQL 2000 SP4 for all three virtual
instances (SQL 1, 2 and 3)

Node 3: Is correctly configured to run SQL 2000 SP4 for SQL Instances 2
and
3, but not Instance 1

Node 4: Is correctly configured to run SQL 2000 SP4 for only SQL Instance
3
(it will run SQL Instances 1 and 2, but only at RTM binaries.


Like I mentioned, I am home sick today so I don't have all the log files
with me. I will try to post them tomorrow.

Regards,

hmscott



.

Relevant Pages

  • Re: Create SQL cluster on 2003
    ... Since you are in a cluster configuration, ... Microsoft SQL Server MVP ... We only about about 20 databases - all ... Move Databases Between Computers That Are Running SQL Server ...
    (microsoft.public.sqlserver.clustering)
  • Re: Unable to Apply SP4 to SQL 2000 Cluster (new Node)
    ... While attempting to install SP4, it comes up immediately after I ... to be on one of the nodes that can bring SQL Server online and apply SP3a, ...
    (microsoft.public.sqlserver.clustering)
  • Re: SQL Server 2005 Cluster Setup Quiz
    ... the first instance on the cluster. ... SQL Server 2005 failover cluster has been installed? ... drive of a SQL Server 2005 failover cluster installation? ...
    (microsoft.public.sqlserver.clustering)
  • Re: A complicate cluster configuration...
    ... The SQL Server 2000 EE will be running just in one node, ... It is possible that an operator can manage the cluster from a terminal ... > and SFTPs them to a remote machine. ...
    (microsoft.public.windows.server.clustering)
  • Re: SQL Server 2005 Cluster Setup Quiz
    ... I did test and it did not install the client tools. ... http://www.clusterhelp.com - Cluster Training ... Microsoft SQL Server MVP ... Provide a template on how to read SQL Server 2005 setup log files. ...
    (microsoft.public.sqlserver.clustering)