Cluster name with multiple IP addresses?

Can multiple IP addresses be assigned to a cluster and/or virtual name?
I have a 2-node Active/Passive cluster on a switched (not routed) network.
Each node has 3 NICs, each on its own subnet;
a) public/client access - 192.168.10.0
b) private/cluster traffic - 10.0.0.0
c) data backup network - 192.168.20.0
The cluster name "SQLCLST" and the virtual machine name "SQLSVR" have been
defined with IP addresses on the public/client network.
My backup software is cluster aware but because the traffic is separated via
switched network, I need the cluster and/or virtual machine name to respond
to IP addresses on two different subnets. Is this possible? Did this make
sense? ;)
TIA
It is not an uncommon request, nor is it too difficult. Once you have the
NICs and networks defined in the cluser configuration, you run the SQL setup
wizard to add the networks to SQL Server. To use them effectively, look
into connection-specific suffixes. That way you can force traffic to a
particular network.
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP
"Structured Chaos" <jeffery_tyree@.yahoo.com> wrote in message
news:%23Y%23s8Mv0HHA.5380@.TK2MSFTNGP04.phx.gbl...
> Can multiple IP addresses be assigned to a cluster and/or virtual name?
> I have a 2-node Active/Passive cluster on a switched (not routed) network.
> Each node has 3 NICs, each on its own subnet;
> a) public/client access - 192.168.10.0
> b) private/cluster traffic - 10.0.0.0
> c) data backup network - 192.168.20.0
> The cluster name "SQLCLST" and the virtual machine name "SQLSVR" have been
> defined with IP addresses on the public/client network.
> My backup software is cluster aware but because the traffic is separated
> via switched network, I need the cluster and/or virtual machine name to
> respond to IP addresses on two different subnets. Is this possible? Did
> this make sense? ;)
> TIA
>
|||Thanks for the quick info. Apparently it is a little more difficult than
that, heheh
I did not see any dialog/screen that allowed for the addition of new
network - only to modify node membership.
I have a ticket opened with MS. They have escalated it stating that they
don't even know if SQL will support this configuration and are attempting to
duplicate it in their labs. The MS SQL Data Engine Team manager stated they
may even have to look at the code...
Should you happen to have any detailed steps on making this work, I'd
appreciate it and apparently MS would too! LOL
"Geoff N. Hiten" <SQLCraftsman@.gmail.com> wrote in message
news:OQYwDPx0HHA.1184@.TK2MSFTNGP04.phx.gbl...
> It is not an uncommon request, nor is it too difficult. Once you have the
> NICs and networks defined in the cluser configuration, you run the SQL
> setup wizard to add the networks to SQL Server. To use them effectively,
> look into connection-specific suffixes. That way you can force traffic to
> a particular network.
> --
> Geoff N. Hiten
> Senior SQL Infrastructure Consultant
> Microsoft SQL Server MVP
>
>
> "Structured Chaos" <jeffery_tyree@.yahoo.com> wrote in message
> news:%23Y%23s8Mv0HHA.5380@.TK2MSFTNGP04.phx.gbl...
>
|||OK, you are not trying to stack multiple IP addresses on a single NIC, you
are using a unique IP address for each NIC. That isn't a problem.
The problem comes from having the same DNS alias (network name) respond on
two different subnets. That is a networking issue, not a clustering issue.
I would use connection-specific domain suffixes to differentiate the
networks. Also, I would use DNS SRV records to abstract the names. I.E.
MySQLCluster points to the IP and port of the normal public NIC.
MySQLClusterBackup points to the IP and port on the backup LAN.
I have used this setup in the past and it works just fine.
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP
"Structured Chaos" <jeffery_tyree@.yahoo.com> wrote in message
news:Ooh45iT2HHA.2752@.TK2MSFTNGP06.phx.gbl...
> Thanks for the quick info. Apparently it is a little more difficult than
> that, heheh
> I did not see any dialog/screen that allowed for the addition of new
> network - only to modify node membership.
> I have a ticket opened with MS. They have escalated it stating that they
> don't even know if SQL will support this configuration and are attempting
> to duplicate it in their labs. The MS SQL Data Engine Team manager stated
> they may even have to look at the code...
> Should you happen to have any detailed steps on making this work, I'd
> appreciate it and apparently MS would too! LOL
>
> "Geoff N. Hiten" <SQLCraftsman@.gmail.com> wrote in message
> news:OQYwDPx0HHA.1184@.TK2MSFTNGP04.phx.gbl...
>

Cluster IP address change and delay

Hi all,

I have recently changed IP addresses on my MS SQL cluster (the new IP
addresses belongs to a different network). Both physical node's IP
addresses and resource's IP addresses have been changed.

Everything seems to work fine. The cluster is up and responding
normally. The resources failover to another node when needed correctly.
However, I have noticed that the time it takes for the resource to
failover to another node is much longer than in the past (before IP
addresses change). I have noticed that "Network Name" is the resource
type that causes the delay. Could you please advise if there is
something that I overlooked and needs to be changed.updated as well, so
that groups fail over faster?

Thanks,
AleuHi

<aleu@.vp.plwrote in message news:f34s2m$b1q$2@.news.onet.pl...

Quote:

Originally Posted by

Hi all,
>
I have recently changed IP addresses on my MS SQL cluster (the new IP
addresses belongs to a different network). Both physical node's IP
addresses and resource's IP addresses have been changed.
>
Everything seems to work fine. The cluster is up and responding normally.
The resources failover to another node when needed correctly. However, I
have noticed that the time it takes for the resource to failover to
another node is much longer than in the past (before IP addresses change).
I have noticed that "Network Name" is the resource type that causes the
delay. Could you please advise if there is something that I overlooked and
needs to be changed.updated as well, so that groups fail over faster?
>
Thanks,
Aleu

I assume that you have followed http://support.microsoft.com/kb/244980 or
http://msdn2.microsoft.com/en-us/library/ms190460.aspx? Which version of SQL
Server is this?

When did you test the failover as it will take time to propogate the IP
address change. Have you tried to ping each of the servers or resolve the
names in the DNS?

John|||John Bell wrote:

Quote:

Originally Posted by

I assume that you have followed http://support.microsoft.com/kb/244980 or
http://msdn2.microsoft.com/en-us/library/ms190460.aspx? Which version of SQL
Server is this?

John, thanks for your response. Yes I did follow MS knowledge base
articles. I have MS SQL server 2005 x64.

Quote:

Originally Posted by

When did you test the failover as it will take time to propogate the IP
address change. Have you tried to ping each of the servers or resolve the
names in the DNS?

I did the ping commands. Well, I have tested the failover many hours
after that...

Any idea why this is so slow?

Thanks,
Aleu|||Hi

On May 28, 2:17 pm, "a...@.vp.pl" <a...@.vp.plwrote:

Quote:

Originally Posted by

John Bell wrote:

Quote:

Originally Posted by

I assume that you have followedhttp://support.microsoft.com/kb/244980or
http://msdn2.microsoft.com/en-us/li...0460.aspx?Which version of SQL
Server is this?

>
John, thanks for your response. Yes I did follow MS knowledge base
articles. I have MS SQL server 2005 x64.
>

Quote:

Originally Posted by

When did you test the failover as it will take time to propogate the IP
address change. Have you tried to ping each of the servers or resolve the
names in the DNS?

>
I did the ping commands. Well, I have tested the failover many hours
after that...
>
Any idea why this is so slow?
>
Thanks,
Aleu

I can't think of any other reasons why this should be an issue if your
DNS and routing are working fine. You may want to try changing it back
and seeing if the time improves.

John|||John Bell wrote:

Quote:

Originally Posted by

>
I can't think of any other reasons why this should be an issue if your
DNS and routing are working fine. You may want to try changing it back
and seeing if the time improves.

Unfortunately, I cannot change it back. Not only the IP addresses of
cluster resources have changed but also, cluster node's IP addresses
(moved to a different network). The "SQL IP address 1" becomes active
very fast (in around 6 seconds), however, "SQL Network Name" takes a few
good minutes (3-4) to change status from "online pending" to "online".

I have checked system logs and found few relevant things:

1. The configuration of the AdminConnection\TCP protocol in the SQL
instance VALIDATION is not valid.

2. Database mirroring connection error 4 'An error occurred while
receiving data: '10053(An established connection was aborted by the
software in your host machine.)'.' for 'TCP://hostname.x.x:2343'.

3. Database mirroring connection error 4 'An error occurred while
receiving data: '64(The specified network name is no longer available.)'.'

4. Configuration option 'Agent XPs' changed from 1 to 0. Run the
RECONFIGURE statement to install.

I do not know whether they are of any use, but (1) seems to be relevant.
What does it mean? Where can I check/change settings of
AdminConnection\TCP protocol?

What is the best place to look for the error logs? Does MS SQL server
store its error logs in some specific files? If so, what is the filename
and it's location?

Sorry for many questions but I am still newbie in this thing.

Thanks,
Aleu|||Hi Aleu

On May 28, 5:55 pm, "a...@.vp.pl" <a...@.vp.plwrote:

Quote:

Originally Posted by

John Bell wrote:
>

Quote:

Originally Posted by

I can't think of any other reasons why this should be an issue if your
DNS and routing are working fine. You may want to try changing it back
and seeing if the time improves.

>
Unfortunately, I cannot change it back. Not only the IP addresses of
cluster resources have changed but also, cluster node's IP addresses
(moved to a different network). The "SQL IP address 1" becomes active
very fast (in around 6 seconds), however, "SQL Network Name" takes a few
good minutes (3-4) to change status from "online pending" to "online".
>
I have checked system logs and found few relevant things:
>
1. The configuration of theAdminConnection\TCPprotocol in the SQL
instance VALIDATION is not valid.
>
2. Database mirroring connection error 4 'An error occurred while
receiving data: '10053(An established connection was aborted by the
software in your host machine.)'.' for 'TCP://hostname.x.x:2343'.
>
3. Database mirroring connection error 4 'An error occurred while
receiving data: '64(The specified network name is no longer available.)'.'
>
4. Configuration option 'Agent XPs' changed from 1 to 0. Run the
RECONFIGURE statement to install.
>
I do not know whether they are of any use, but (1) seems to be relevant.
What does it mean? Where can I check/change settings ofAdminConnection\TCPprotocol?
>
What is the best place to look for the error logs? Does MS SQL server
store its error logs in some specific files? If so, what is the filename
and it's location?
>
Sorry for many questions but I am still newbie in this thing.
>
Thanks,
Aleu

There are quite a few posts if you look for "AdminConnection\TCP"
http://groups.google.com/groups/sea...ection%5CTCP%22
they seems to suggest networking issues (incorrect aliases, firewall
issues etc) or possibly service account permissions.

I assume that the error messages regarding mirroring are related to
the period when you are switching, therefore may be expected.

John

Cluster Addresses

Hi,
I am a newbie to Clustering so please bear with me.
We have a db environment set up with Win 2003 and SQL 2000 connecting to a
single disk array.
The question I have is, what is the difference between the Cluster IP
Address and the SQL IP address.
What I really want to know is what is the purpose for the two different
addresses?
Thanks for any help.
JD
When you create a cluster, you start working with Virtual Servers. The
cluster itself becomes a Virtual server, as does each SQL instance you
install. Each virtual server consists of a minimum of a physical disk
resource, an IP address, and a network name. Of course, each virtual server
has other resources to make it actually do something useful. So, the
Cluster IP address is used to connect to Cluster control resources and the
SQL IP address is used to connect to SQL service resources. These addresses
are in addition to the IP addresses of the host nodes.
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"jimmy" <jimmy@.discussions.microsoft.com> wrote in message
news:DC5A7EF8-C516-411B-823D-BA36A6E24A85@.microsoft.com...
> Hi,
> I am a newbie to Clustering so please bear with me.
> We have a db environment set up with Win 2003 and SQL 2000 connecting to a
> single disk array.
> The question I have is, what is the difference between the Cluster IP
> Address and the SQL IP address.
> What I really want to know is what is the purpose for the two different
> addresses?
> Thanks for any help.
> JD
|||Hi Geoff,
Thanks for the explanation. We have a single application that points at the
database, so should we use the SQL IP Address to point the application at the
db server?
So from your explanaition if a file share was one of the Cluster resources
and you wanted to acces it. This is when you would use the Cluster address
to access the file share. Like so:
\\ClusterIP\FileShare
Thanks Again
JD
"Geoff N. Hiten" wrote:

> When you create a cluster, you start working with Virtual Servers. The
> cluster itself becomes a Virtual server, as does each SQL instance you
> install. Each virtual server consists of a minimum of a physical disk
> resource, an IP address, and a network name. Of course, each virtual server
> has other resources to make it actually do something useful. So, the
> Cluster IP address is used to connect to Cluster control resources and the
> SQL IP address is used to connect to SQL service resources. These addresses
> are in addition to the IP addresses of the host nodes.
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
> I support the Professional Association for SQL Server
> www.sqlpass.org
> "jimmy" <jimmy@.discussions.microsoft.com> wrote in message
> news:DC5A7EF8-C516-411B-823D-BA36A6E24A85@.microsoft.com...
>
>
|||Yes, you use the SQL vurtial server IP address as the connection target.
Personally, I prefer to use the network name. That way, I can play with
Aliases and DNS records during disasters.
You shouldn't use the cluster resource group and quorum disk for anything
extra. You should create another virtual server and add the file service to
that resource group. \\FileShareVirtualServerName\ShareName
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"jimmy" <jimmy@.discussions.microsoft.com> wrote in message
news:641B055A-4878-4E12-AE77-DE4BFDEF3718@.microsoft.com...
> Hi Geoff,
> Thanks for the explanation. We have a single application that points at
the
> database, so should we use the SQL IP Address to point the application at
the
> db server?
> So from your explanaition if a file share was one of the Cluster resources
> and you wanted to acces it. This is when you would use the Cluster
address[vbcol=seagreen]
> to access the file share. Like so:
> \\ClusterIP\FileShare
> Thanks Again
> JD
> "Geoff N. Hiten" wrote:
server[vbcol=seagreen]
the[vbcol=seagreen]
addresses[vbcol=seagreen]
to a[vbcol=seagreen]
different[vbcol=seagreen]
|||Hi Geoff,
Thats great.
Thanks
Jamie.
"Geoff N. Hiten" wrote:

> Yes, you use the SQL vurtial server IP address as the connection target.
> Personally, I prefer to use the network name. That way, I can play with
> Aliases and DNS records during disasters.
> You shouldn't use the cluster resource group and quorum disk for anything
> extra. You should create another virtual server and add the file service to
> that resource group. \\FileShareVirtualServerName\ShareName
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
> I support the Professional Association for SQL Server
> www.sqlpass.org
> "jimmy" <jimmy@.discussions.microsoft.com> wrote in message
> news:641B055A-4878-4E12-AE77-DE4BFDEF3718@.microsoft.com...
> the
> the
> address
> server
> the
> addresses
> to a
> different
>
>
|||So...
Are there any restrictions on the Cluster IP and SQL IP?
If I have an internal and external NIC, should one go on each? Can they have neighboring IPs?
The reason I ask is that I set up a new cluster with the Cluster IP on the external NIC and the SQL IP on the internal NIC. At this point everything worked fine. However the requirements changed and I changed the SQL IP from the internal NIC to the external one - with an IP contigious to the Cluster IP. Ever since, I can not get my application to connect over TCP. Sometimes the sql services will not come online either until I change the IPs back to the original settings.
I have changed the IPs both in Cluster Admin and using the SQL installer as indicated in this KB article: http://support.microsoft.com/kb/244980.
Any thoughts?

Cluster access using IP

I joined a company as a DBA
the company has cluster servers set up.
I was given the 3 ip addresses. i for the cluster and 2 for for the ones the
cluster is made up of. and my login is set up to use SQL Server
Authentication.
How do i connect to the server. I do not know if they are active/active or
active/passive. I do not have any rights on the local boxes for me to telnet
them.
What is the way to connect to the database
what is the default instance that SQl server creates.
thanks a lots in advance
Connect to the address of the SQL Virtual Server, not just the cluster
address or the address of either machine (node).
Cheers,
Rod
MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering
http://www.msmvps.com/clustering - Blog
"Nik" <patelnikhil30@.hotmail.com> wrote in message
news:%23%230sxSelEHA.3156@.TK2MSFTNGP12.phx.gbl...
>I joined a company as a DBA
> the company has cluster servers set up.
> I was given the 3 ip addresses. i for the cluster and 2 for for the ones
> the
> cluster is made up of. and my login is set up to use SQL Server
> Authentication.
> How do i connect to the server. I do not know if they are active/active or
> active/passive. I do not have any rights on the local boxes for me to
> telnet
> them.
> What is the way to connect to the database
> what is the default instance that SQl server creates.
> thanks a lots in advance
>