Cluster Upgrade SP2 2005

Have recently performed an in-place upgrade of a 2 instance sql cluster from 2000 to 2005. The upgrade process went relatively smoothly. However, I am running into problems installed sp2. The first problem that I run into is that I have resrcmon.exe locked. I can click next, which is fine, but once I get past that point I get stuck during the installation process. I am updating DatabaseServices and Integration Services. It gets stuck on 'Awaiting first complete passive cluster node SQL-Test'. Does anyone have any ideas?
Thanks,
TimFor what it's worth I'm experiencing the same issue as described above. I'm using VMWare Server and have a Windows 2003 Enterprise Edition cluster with SQL Server 2005 Standard Edition clustered on top. This works well but in trying to apply SP2 it is stuck "Awaiting first complete passive cluster node". Slight difference from that described above, I had no locked files.

Any info that anyone could provide would be greatly appreciated.

Cheers,

Ian
|||Unbelievable. Typically just as I finish writing the above, the process moved on. However it failed at the "Database Services" product. Everything else seemed to go okay but I'll have to check and will come back with any info I find.

Cheers,

Ian
|||I got my issue resolved...after talking to MS Support for half of a day. There are a lot of things that you have to make sure is just right. For example, when you install the services, it is easier to have the first node control both instances. From there, make sure you install the client tools on the default instance first. You also cannot have remote connections to the 2nd node in the cluster (will cause it to fail). Mine failed on installed the database services as well. Another thing we had to change was that when the network guys setup the cluster, the distributed transaction coordinator wasn't created as a clustered resource...it needs to be. let me know if this helps any.
Tim|||Tim,

Did you manage to continue with resrcmon.exe locked or did you find a way to stop it (it's part of the cluster so, short of stopping the cluster services I wasn't sure how else to stop it)?

We've got the DTC in place in our production cluster but I forgot to add it with my virtual one. I've done this now and will give it another go shortly.

Cheers,

Ian
|||At first I seen that resrcmon.exe was locked, and it did concern me. However, I don't think that it really matters if it is locked or not. If it is locked, you can still go through with the install, but it will prompt you to reboot later in the process. Try to kill all remote connections to either cluster node, place both instances on one node, and try to install that way. That will likely unlock resrcmon.|||I actually only have one instance, it is on the node from which I'm running the SP. WRT remote connections, I can't see anything apart from the clustering services that would be connecting remotely to the second node.

That said, I've only got resrcmon.exe locked now so am gonna proceed with the SP and see where I get - beauty of VMs is that I can just roll 'em back any time I like :-)
|||

I had the exact same issue. Installing the client tools on both nodes, and then moving the inst1 over to the passive node and running the SP2 from there seemed to do the trick.

|||I've finally had some level of success with this on my virtual cluster. But it was kinda forceful. I basically downed node 2, installed the SP onto node 1 (twice as it had "reboot required" on the SQL Server engine the first time). Then brought up node 2, failed over and downed node 1. Did the same to node 2 (twice) and brought node 1 back up. Testing seems to confirm this has gone okay. However, as we're in the process of trying to get a test cluster (for exactly this kind of thing - good as VMWare is, it's not the real thing) I'm not going to apply this to production until that's happened.

Cheers,

Ian
|||

Hi Ian,

Please help - when you downed the one node - was it shutdown or did you stop cluster service on the node or did you pause the node?

Thanks in advance :-)

|||

Hi there,

Basically I did the following.

Node 1 was up, node 2 was shutdown completely. I upgraded node 1 to SP2, then brought node 2 back online and failed over by shutting down node 1. I then upgraded node 2 to SP2, brought node 1 back online, failed over manually to confirm all as expected and left it at that. Bear in mind this was on virtual machines so is not a perfect test. I've yet to do this on our production cluster and am currently trying to get a proper physical test cluster in place to run through this for real.

Hope this helps!

Cheers,

Ian

Cluster Upgrade SP2 2005

Have recently performed an in-place upgrade of a 2 instance sql cluster from 2000 to 2005. The upgrade process went relatively smoothly. However, I am running into problems installed sp2. The first problem that I run into is that I have resrcmon.exe locked. I can click next, which is fine, but once I get past that point I get stuck during the installation process. I am updating DatabaseServices and Integration Services. It gets stuck on 'Awaiting first complete passive cluster node SQL-Test'. Does anyone have any ideas?
Thanks,
TimFor what it's worth I'm experiencing the same issue as described above. I'm using VMWare Server and have a Windows 2003 Enterprise Edition cluster with SQL Server 2005 Standard Edition clustered on top. This works well but in trying to apply SP2 it is stuck "Awaiting first complete passive cluster node". Slight difference from that described above, I had no locked files.

Any info that anyone could provide would be greatly appreciated.

Cheers,

Ian
|||Unbelievable. Typically just as I finish writing the above, the process moved on. However it failed at the "Database Services" product. Everything else seemed to go okay but I'll have to check and will come back with any info I find.

Cheers,

Ian
|||I got my issue resolved...after talking to MS Support for half of a day. There are a lot of things that you have to make sure is just right. For example, when you install the services, it is easier to have the first node control both instances. From there, make sure you install the client tools on the default instance first. You also cannot have remote connections to the 2nd node in the cluster (will cause it to fail). Mine failed on installed the database services as well. Another thing we had to change was that when the network guys setup the cluster, the distributed transaction coordinator wasn't created as a clustered resource...it needs to be. let me know if this helps any.
Tim|||Tim,

Did you manage to continue with resrcmon.exe locked or did you find a way to stop it (it's part of the cluster so, short of stopping the cluster services I wasn't sure how else to stop it)?

We've got the DTC in place in our production cluster but I forgot to add it with my virtual one. I've done this now and will give it another go shortly.

Cheers,

Ian
|||At first I seen that resrcmon.exe was locked, and it did concern me. However, I don't think that it really matters if it is locked or not. If it is locked, you can still go through with the install, but it will prompt you to reboot later in the process. Try to kill all remote connections to either cluster node, place both instances on one node, and try to install that way. That will likely unlock resrcmon.|||I actually only have one instance, it is on the node from which I'm running the SP. WRT remote connections, I can't see anything apart from the clustering services that would be connecting remotely to the second node.

That said, I've only got resrcmon.exe locked now so am gonna proceed with the SP and see where I get - beauty of VMs is that I can just roll 'em back any time I like :-)
|||

I had the exact same issue. Installing the client tools on both nodes, and then moving the inst1 over to the passive node and running the SP2 from there seemed to do the trick.

|||I've finally had some level of success with this on my virtual cluster. But it was kinda forceful. I basically downed node 2, installed the SP onto node 1 (twice as it had "reboot required" on the SQL Server engine the first time). Then brought up node 2, failed over and downed node 1. Did the same to node 2 (twice) and brought node 1 back up. Testing seems to confirm this has gone okay. However, as we're in the process of trying to get a test cluster (for exactly this kind of thing - good as VMWare is, it's not the real thing) I'm not going to apply this to production until that's happened.

Cheers,

Ian
|||

Hi Ian,

Please help - when you downed the one node - was it shutdown or did you stop cluster service on the node or did you pause the node?

Thanks in advance :-)

|||

Hi there,

Basically I did the following.

Node 1 was up, node 2 was shutdown completely. I upgraded node 1 to SP2, then brought node 2 back online and failed over by shutting down node 1. I then upgraded node 2 to SP2, brought node 1 back online, failed over manually to confirm all as expected and left it at that. Bear in mind this was on virtual machines so is not a perfect test. I've yet to do this on our production cluster and am currently trying to get a proper physical test cluster in place to run through this for real.

Hope this helps!

Cheers,

Ian

sqlsql

Cluster Upgrade SP2 2005

Have recently performed an in-place upgrade of a 2 instance sql cluster from 2000 to 2005. The upgrade process went relatively smoothly. However, I am running into problems installed sp2. The first problem that I run into is that I have resrcmon.exe locked. I can click next, which is fine, but once I get past that point I get stuck during the installation process. I am updating DatabaseServices and Integration Services. It gets stuck on 'Awaiting first complete passive cluster node SQL-Test'. Does anyone have any ideas?
Thanks,
TimFor what it's worth I'm experiencing the same issue as described above. I'm using VMWare Server and have a Windows 2003 Enterprise Edition cluster with SQL Server 2005 Standard Edition clustered on top. This works well but in trying to apply SP2 it is stuck "Awaiting first complete passive cluster node". Slight difference from that described above, I had no locked files.

Any info that anyone could provide would be greatly appreciated.

Cheers,

Ian
|||Unbelievable. Typically just as I finish writing the above, the process moved on. However it failed at the "Database Services" product. Everything else seemed to go okay but I'll have to check and will come back with any info I find.

Cheers,

Ian
|||I got my issue resolved...after talking to MS Support for half of a day. There are a lot of things that you have to make sure is just right. For example, when you install the services, it is easier to have the first node control both instances. From there, make sure you install the client tools on the default instance first. You also cannot have remote connections to the 2nd node in the cluster (will cause it to fail). Mine failed on installed the database services as well. Another thing we had to change was that when the network guys setup the cluster, the distributed transaction coordinator wasn't created as a clustered resource...it needs to be. let me know if this helps any.
Tim|||Tim,

Did you manage to continue with resrcmon.exe locked or did you find a way to stop it (it's part of the cluster so, short of stopping the cluster services I wasn't sure how else to stop it)?

We've got the DTC in place in our production cluster but I forgot to add it with my virtual one. I've done this now and will give it another go shortly.

Cheers,

Ian
|||At first I seen that resrcmon.exe was locked, and it did concern me. However, I don't think that it really matters if it is locked or not. If it is locked, you can still go through with the install, but it will prompt you to reboot later in the process. Try to kill all remote connections to either cluster node, place both instances on one node, and try to install that way. That will likely unlock resrcmon.|||I actually only have one instance, it is on the node from which I'm running the SP. WRT remote connections, I can't see anything apart from the clustering services that would be connecting remotely to the second node.

That said, I've only got resrcmon.exe locked now so am gonna proceed with the SP and see where I get - beauty of VMs is that I can just roll 'em back any time I like :-)
|||

I had the exact same issue. Installing the client tools on both nodes, and then moving the inst1 over to the passive node and running the SP2 from there seemed to do the trick.

|||I've finally had some level of success with this on my virtual cluster. But it was kinda forceful. I basically downed node 2, installed the SP onto node 1 (twice as it had "reboot required" on the SQL Server engine the first time). Then brought up node 2, failed over and downed node 1. Did the same to node 2 (twice) and brought node 1 back up. Testing seems to confirm this has gone okay. However, as we're in the process of trying to get a test cluster (for exactly this kind of thing - good as VMWare is, it's not the real thing) I'm not going to apply this to production until that's happened.

Cheers,

Ian
|||

Hi Ian,

Please help - when you downed the one node - was it shutdown or did you stop cluster service on the node or did you pause the node?

Thanks in advance :-)

|||

Hi there,

Basically I did the following.

Node 1 was up, node 2 was shutdown completely. I upgraded node 1 to SP2, then brought node 2 back online and failed over by shutting down node 1. I then upgraded node 2 to SP2, brought node 1 back online, failed over manually to confirm all as expected and left it at that. Bear in mind this was on virtual machines so is not a perfect test. I've yet to do this on our production cluster and am currently trying to get a proper physical test cluster in place to run through this for real.

Hope this helps!

Cheers,

Ian

Cluster backup

Quick question...
2 node active/passive (excuse the terminology...) cluster.
My network admin wants to have my SQL backups (performed via. maint. plan
nightly) stored on the quorum drive.
We are also performing log shipping at 15 minute intervals to an offsite
server.
I think storing the backups on the quorum drive is a bad idea as if the disk
array dies completely, we have no backup.
Also, would that scheme not necessitate adding the quorum drive as a
resource to the SQL Server?
Comments and thoughts please.
Thanks in advance.
B.
You are correct, its a bad idea The Quorum drive would have to be a
dependency for SQL. Which would mean one group for basically everything. If
the backups filled the disk, the cluster would not be happy.
A good idea is use the MP to a disk, and then backup to some kind of tape.
Cheers,
Rod
MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering
http://msmvps.com/clustering - Blog
"Tam O'Shanter" <Tam@.Oshanter.com> wrote in message
news:b6prd.6024491$6p.961500@.news.easynews.com...
> Quick question...
> 2 node active/passive (excuse the terminology...) cluster.
> My network admin wants to have my SQL backups (performed via. maint. plan
> nightly) stored on the quorum drive.
> We are also performing log shipping at 15 minute intervals to an offsite
> server.
> I think storing the backups on the quorum drive is a bad idea as if the
> disk
> array dies completely, we have no backup.
> Also, would that scheme not necessitate adding the quorum drive as a
> resource to the SQL Server?
> Comments and thoughts please.
> Thanks in advance.
> B.
>
>
|||Thanks Rodney,
Validation of what I thought.
Would you suggest backing up to a local disk on one of the nodes? Or rather,
having the MP write to a UNC on another machine in the environment
(management server etc..).
I think the UNC is the way to go, but I'd be interested in your thoughts.
Every morning the bak's would be moved from the backup location to an
external firewire drive and physically removed from the facility. This
setup, in conjunction with a logshp (15 min. intervals) to an offsite box
seems like my best option.
Comments?
Thanks again.
B.
"Rodney R. Fournier [MVP]" <rod@.die.spam.die.nw-america.com> wrote in
message news:ORI$OI%231EHA.1192@.tk2msftngp13.phx.gbl...
> You are correct, its a bad idea The Quorum drive would have to be a
> dependency for SQL. Which would mean one group for basically everything.
If[vbcol=seagreen]
> the backups filled the disk, the cluster would not be happy.
> A good idea is use the MP to a disk, and then backup to some kind of tape.
> Cheers,
> Rod
> MVP - Windows Server - Clustering
> http://www.nw-america.com - Clustering
> http://msmvps.com/clustering - Blog
> "Tam O'Shanter" <Tam@.Oshanter.com> wrote in message
> news:b6prd.6024491$6p.961500@.news.easynews.com...
plan
>
|||Tam, you have a problem. You can't backup SQL using the MP to a UNC, though
the last time I tried was years ago. I also don't like the idea of local
node backups. Can you add a disk to the array? Or does the current array
have space already? Remember with the MP you can delete old backups to keep
the space minimal.
Rod
"Tam O'Shanter" <Tam@.Oshanter.com> wrote in message
news:kXprd.3931798$yk.598061@.news.easynews.com...
> Thanks Rodney,
> Validation of what I thought.
> Would you suggest backing up to a local disk on one of the nodes? Or
> rather,
> having the MP write to a UNC on another machine in the environment
> (management server etc..).
> I think the UNC is the way to go, but I'd be interested in your thoughts.
> Every morning the bak's would be moved from the backup location to an
> external firewire drive and physically removed from the facility. This
> setup, in conjunction with a logshp (15 min. intervals) to an offsite box
> seems like my best option.
> Comments?
> Thanks again.
> B.
> "Rodney R. Fournier [MVP]" <rod@.die.spam.die.nw-america.com> wrote in
> message news:ORI$OI%231EHA.1192@.tk2msftngp13.phx.gbl...
> If
> plan
>
|||Yes, you can backup to a UNC. I do it all the time. There is even a
community KB article on how to do backup to a UNC share. Personally, I
think it is one of the best ways to back up a system.
HowTo: Backup to UNC name using Database Maintenance Wizard
http://support.microsoft.com/?kbid=555128
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"Rodney R. Fournier [MVP]" <rod@.die.spam.die.nw-america.com> wrote in
message news:u9yvudB2EHA.2568@.TK2MSFTNGP10.phx.gbl...
> Tam, you have a problem. You can't backup SQL using the MP to a UNC,
though
> the last time I tried was years ago. I also don't like the idea of local
> node backups. Can you add a disk to the array? Or does the current array
> have space already? Remember with the MP you can delete old backups to
keep[vbcol=seagreen]
> the space minimal.
> Rod
> "Tam O'Shanter" <Tam@.Oshanter.com> wrote in message
> news:kXprd.3931798$yk.598061@.news.easynews.com...
thoughts.[vbcol=seagreen]
box[vbcol=seagreen]
everything.[vbcol=seagreen]
the
>
|||Excellent. I will file that away for future use Maybe it was 7.0 or 6.5
or 4.21 or 4.2, I have been using SQL for a long time. What I don't like
about UNC is that any network glitch or outage can potentially kill the
backup.
Cheers,
Rod
MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering
http://msmvps.com/clustering - Blog
"Geoff N. Hiten" <SRDBA@.Careerbuilder.com> wrote in message
news:O845A2H2EHA.2568@.TK2MSFTNGP11.phx.gbl...
> Yes, you can backup to a UNC. I do it all the time. There is even a
> community KB article on how to do backup to a UNC share. Personally, I
> think it is one of the best ways to back up a system.
> HowTo: Backup to UNC name using Database Maintenance Wizard
> http://support.microsoft.com/?kbid=555128
>
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
> I support the Professional Association for SQL Server
> www.sqlpass.org
> "Rodney R. Fournier [MVP]" <rod@.die.spam.die.nw-america.com> wrote in
> message news:u9yvudB2EHA.2568@.TK2MSFTNGP10.phx.gbl...
> though
> keep
> thoughts.
> box
> everything.
> the
>
|||Yep. You can also easily overwhelm a network. I use a dedicated link with
a connection-specific domain suffix to segment the backup traffic from the
normal SQL traffic. I also tune the file share to maximize write
capability. I just upgraded the backup host server (which is also my tape
host server) to handle the load.
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"Rodney R. Fournier [MVP]" <rod@.die.spam.die.nw-america.com> wrote in
message news:%23kJFlsI2EHA.2572@.tk2msftngp13.phx.gbl...[vbcol=seagreen]
> Excellent. I will file that away for future use Maybe it was 7.0 or 6.5
> or 4.21 or 4.2, I have been using SQL for a long time. What I don't like
> about UNC is that any network glitch or outage can potentially kill the
> backup.
> Cheers,
> Rod
> MVP - Windows Server - Clustering
> http://www.nw-america.com - Clustering
> http://msmvps.com/clustering - Blog
> "Geoff N. Hiten" <SRDBA@.Careerbuilder.com> wrote in message
> news:O845A2H2EHA.2568@.TK2MSFTNGP11.phx.gbl...
local[vbcol=seagreen]
array[vbcol=seagreen]
This[vbcol=seagreen]
a[vbcol=seagreen]
maint.[vbcol=seagreen]
if[vbcol=seagreen]
a
>
|||We use the Tivoli suite of products...Tivoli storage manager for the OS, and
Tivoli data protection for SQL. We cart all back ups off the servers in case
there is a drive failure. As long as your cluster is active / passive, it's
not too intense to set up and use, but active / active gets a bit trickier.
both nodes need to be able to authenticate to the back up server as each othr
on active / active....we have 2 such clusters running this way, and 5 more
active/ passives.
"Geoff N. Hiten" wrote:

> Yep. You can also easily overwhelm a network. I use a dedicated link with
> a connection-specific domain suffix to segment the backup traffic from the
> normal SQL traffic. I also tune the file share to maximize write
> capability. I just upgraded the backup host server (which is also my tape
> host server) to handle the load.
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
> I support the Professional Association for SQL Server
> www.sqlpass.org
> "Rodney R. Fournier [MVP]" <rod@.die.spam.die.nw-america.com> wrote in
> message news:%23kJFlsI2EHA.2572@.tk2msftngp13.phx.gbl...
> local
> array
> This
> a
> maint.
> if
> a
>
>
sqlsql