DB Engine :: Defining MAX And MIN Server Memory Settings On 2 Node Cluster
Sep 29, 2015
On first Node A: The server has 16 GB of physical RAM.
On second Node B: The server has 10 GB of physical RAM.
Now, this being Active Active, Node A can be clustered on failure onto Node B..Now reporting server is configured under these two nodes, with defined MAX and MIN server memory of 12 as MAX and 0 is min IN GB.Now with this setting on SQL whenever the cluster moves, such config make OS goes low on node for 10 GB.I am only left with option of switching this MAX and MIN to a default setting or is there any other alternative such as script which can change this setting accordingly when cluster moves to respective server.
We have a new set up on VM to run an application running 24*7 (migrated from SQL server 2008R2) with below configs:
1. OS- Windows server 2012 Standard 64 bit hosted on Virtual Machine
2. Memory 16 GB and Cores =4 with 2.4GhZ processor
3. SQL server 2012 SP2 , 64 bit Standard edition.
4. Total size of databases as of now 15 GB with biggest being 5 GB.
How should i go around in setting the MAX and MIN server memory settings. I have this set up for many of SQL 2005 and 2008R2 servers, but for 2012 i heard that things has slightly changed.
How should i start analyzing and setting the right value of this MAX and MIN?
I have an Active/Active/Passive cluster with 64GB RAM on each node running SQL 2000 EE, AWE is enabled as well as the PAE switch, all is dandy with that.
Question: Should I configure each SQL Instance to have only a max mem usage of 32GB in the event both failover to the same node ? or will the memory allocation be handled without any issue if each node is configured to use 64GB ?
So I'm in the middle of building SQL Server 2005 on my new cluster hardware. I've got all the goodies this time around -4x PowerEdge 68004x dual core x64 procs in each box32GB ram in each boxWindows Server 2k3 R2 Enterprise x64 SP2SQL 2005 Enterprise x64 SP2Active Active Active Passive cluster (4 boxes, 3 instances)dedicated data, log, and tempDB 4Gb 15k fiber channel SAN Volumes for each instanceNow, this is the first x64 box I've configured, so I'm looking for some optimization tips for a couple things.First one's memory. On our 32bit systems, I always added the /PAE switch to the boot.ini file and enable AWE in SQL Server. From what I'm reading, that's no longer needed with the enhanced memory addressing. Is that the case? Is there an MS best practices KB doc in regards to configuration? I can't find one.Also, should I set anything special pagefile-wise? I know the old mantra is 1.5x system memory, but that would create a 48GB page file. Our current cluster has 6GB of ram dedicated to each instance (2k) and Perfmon tells me it's barely touching the pagefile. I'm thinking of just leaving it default, but I want to know if there is a best practice that I'm missing. Again, I can't find an MS doc that addresses this particular scenario.
We are running with a 2 node windows cluster having three SQL instances on it.
OS: Windows server 2008R2 SP1 SQL : SQL server 2008R2 (10.50.6529)
Currently both nodes have 256 GB or memory and we are having multiple auto failover for resources. What will be the best practice for OS memory reservation (OS+tools) so that we can set SQL max memory settings accordingly?
Could someone give me an ideal of what it takes to upgrade the memory in my cluster. Is it as easy has upgrading the passive node, switching the nodes, then upgrading the other server. And do I need to re-configure the windows cluster. And will both server need to be down at the same time at any point.
Or is their more to the process? Thanks for any assistance.
But I'm not sure if I have to install SQL Server first on node 2, then add it to the cluster. Or does adding it to the cluster also install the software?
I'm contemplating running two availability groups on a two node WSFC. The WSFC is setup with a file share witness (i.e. no shared storage). Can I safely run 1 AG on one primary node, and the other AG on the other node (as primary). Each AG would have replicas on the passive node. This would effectively allow both servers to be in use at the same time. In a failover event, I understand that both workloads would transfer to a single server - so the box needs to be sized appropriately.
We are in the process of building a 3 node SQL Server Cluster (Server 2012/ SQL Server 2012), and we have configured the quorum so that all 3 nodes have a vote (no file share witness as we already have an odd number of nodes).
As I understand it, this should allow the cluster to run as long as 2 of the nodes remain online.
However, the validation report states that 2 node failures would be acceptable and, when we tested this by powering off two of the nodes, the cluster did indeed continue to run on a single node.
I configure Windows 2003 R2 and SQL 2005 two nodes Cluster. When I move cluster resource from one node to anther node it takes around 30 seconds to become online. So in that time if any query is running it stops responding.
We have been working on a project to upgrade the servers in our 2-node SQL Server environment. I evicted a node after removing it from the instance. We added the new node under a new server name. I then start the Add remove programs, choose to change the SQL 2005 environment, type the virtual server name. Choose to maintain virtual server, pick to add new node. All seems well, I enter all prompted questions, and when the install begins I get the error below.
Product: Microsoft SQL Server 2005 -- Error 1706. An installation package for the product Microsoft SQL Server 2005 cannot be found. Try the installation again using a valid copy of the installation package 'SqlRun_SQL.msi'.
So I copied the SS2K5 Enteerprise Edition software to the local C: drive, point it too the 'SqlRun_SQL.msi' in the setup folder and still get the error.
During the installation of Adding node to a SQL Server failover cluster(On passive node) getting error like.. The MOF compiler could not connect with the WMI server. This is either because of a semantic error such as an incompatibility with the existing WMI repository or an actual error such as the failure of the WMI server to start.We run the below commands but didn’t get any resolution & got the same above error .
1<sup>st</sup> Method…
1. Open console command (Run->CMD with administrator privileges).
2. net stop winmgmt
3. Rename folder %windir%System32WbemRepository to other one, for backup purposes (for example _Repository).
We have SQL Server 2005 Entp 32-bit on Windows Server 2003 Entp.Edtion 64-bit. ( Total memory = 16 Gb)
1) So, in this case , do i need to enable AWE Configuration from SQL Server Side ( 32-bit).
And what settings( AWE, /PAE) i should enable on OS (i.e.) on Windows Server 2003 (64-bit) in boot.ini file. Because right now, i dont see any switches in boot.ini file.
2) And with 16 GB of RAM, what should be the ideal setting for memory ( Min & Max).
We have 2 nodes window Server 2012 R2 and SQL Server 2012 Enterprise Version cluster setup. We can switch roles and Node to one node to another and revert back to previous node with out any issues. But we are facing when one Node is restarted. We could not restart that Node in cluster Service start in Failover cluster Manager. Error Details is displayed as below inside double code."Cluster node NODE1 could not to join the cluster because it failed to communicate over the network with any other node in the cluster. Verify the network connectivity and configuration of any network firewalls."
I checked windows firewall. windows firewall is all of in Node1, Node2, SAN and DC.I have disabled and enabled the Internal and private network of Node 1. I have validated the cluster. it is showing no error though.
Node1: Public IP: 10.10.0.11 SubNet Mask:255.255.255.0 Default Getway: 10.10.0.1 Prefered DNS: 10.10.0.10 (Ip of DNS)
[code]....
Private Network: Not configured.pinging to each other ip is successful from one node to another.
I've been researching AWE to determine if we should enable this for our environment.
Currently we have a quad core box with 4 gb of RAM (VMware). OS: Windows 2003 std, SQL Server 2005 std. 3GB is not set but will be as soon as we can perform maintenance on the server.
I have read mixed feedback on AWE, either it works great or grinds you to a hault. I would assume that the grinding to a hault is due to not setting the min/max values correctly or not enabling the lock page in memory setting.
We only have one instance of SQL on the server and this box won't be used for anything else aside from hosting SQL services. We do plan on running SSRS off of this server as well.
1. Will running SSRS and enabling AWE cause me problems? Will I have to reduce the max setting by the SSRS memory usage or will it share and play nice?
2. How do I go about setting the Max value? Should it be less than the physical RAM in the box? Right now its set to the default of 214748364, even if I don't enable AWE should this default value be changed?
3. It seems that even at idle the SQL server holds a lot of memory and the page file grows. If I restart the process in the morning, memory usage in taskmon is at 600mb or so. By the end of the day, its up around 2gb. How can I track down whats causing this, should this even concern me?
4. The lock Page in memory setting worries me. Everything I've read on this seems to give a warning about serious OS and other program support degradation. In some cases to the point where they have to restore the settings on the server before they can bring it back up. What are your thoughts on this.
I invoke xp_cmdshell proc from inside a stored procedure on a 2-node active/passive SQL 2005 SP2 Standard cluster. Depending on which server the xp_cmdshell gets executed on I need to pass different arguments in the shell command. I thought I could use host_name() function to get the runtime process server, however, I am finding that it's not behaving correctly. In one example I know my active node is server2, but the host_name() function is returning server1. The only thing that I could possible explain this is that the MSDTC cluster group is not always on the same active node as the SQL server group and in the case I am talking about the cluster groups are in this mode (differnet nodes). Does the xp_cmdshell get executed by the SQL active node or the MDTC active node? And what is the best way to find out which server is going to run my xp_cmdshell?
Thanks.
Edit:
Perhaps another by product of this is that if I run select host_name() from the Studio Management query window i get different results depending on which server I am running the Studio Management on. On server1 I get server1 and on server 2 I get server 2, all the while server2 is the active node. I need a different function that will always let me determine the correct server that'll be running the xp_cmdshell...
Edit 2: I guess I could determine the running host inside the command shell itself, but I am curious to see if i can do it (cleaner) from SQL.
I'm having trouble finding the optimum memory settings for SQL Server 2005. I have 4 instances running on a macine with 8 dual-core processors and 18GB of RAM. I have tried the following memory settings so far:
No maximum - one instance used about 12GB of RAM so then the others struggled Maximum of 4GB each (2GB left for Windows) - meant that 3 instances could be using 1GB each and then another at 4GB and needing more whilst 9GB sat unused. Minimum of 2GB each - one instance would use up 10GB and then never give any back to the other instances.
I also find that setting a maximum then just causes a high amount of paging. What I would like to do is have each instance use a minimum amount, say 1GB, and then have each instance use a maximum of 13GB (3GB for other 3 instances running at a minimum level and 2GB for Windows). This 13GB should then be released and allocated to another instance when necessary, assuming it is no longer all being used. I do not want paging to occur if an instance reaches 13GB.
How do I go about configuring SQL Server to behave like this? Is it possible?
My question is, I'd like my actual cluster to have 4 servers. The virtual server that will be running SQL 2005 will only be installed on 2 of those nodes. Will this work or do I have to split my cluster in to 2 to meet the SQL Server 2005 2 node failover limitation? A basic config for the cluster would be like below....
ClusterSQLNode Active SQL 2005 Server ClusterExchNode Active Exchange Server ClusterAppNode Active Application Server ClusterFailoverNode Failover SQL 2005 Failover Exchange Failover Application
Will SQL 2005 install in a 4 server cluster if I only install it on to two of the nodes? Any potential gotchas?
I ran into problem while trying to install SQL Server 2008 to join to another node in a failover cluster ,you will find here the configuration doc and installation details. URL...
When adding a node to a SQL Server 2012 Standard edition cluster, how I do I identify the location for SQL server shared components and the rest of the SQL Server installation binaries?
When adding a node to a SQL Server 2012 Standard edition cluster all the binaries went to the C: drive default location. We put those files on a different drive when installing the first node. What needs to be done so both nodes have the binaries on the same drives and folders?
I have destroyed the cluster in failover cluster manager and then i am trying to remove node from the sql server installation centre.I am facing the cluster node and cluster service verification errors.I am not able to start the cluster service in services as well.
We have (had) an active/active cluster. 2 physical machines, each running their own instance, clustered together. Node1/Ins1 and Node2/Ins2.
Node2 failed and Ins2 failed over to Node1 as it should. Node2 required that we rebuild the server (rebuild = reinstall O/S). Now we need to get Node2 back into the cluster and get Ins2 failed back over to Node2.
Does anyone know, for certain, the correct steps to accomplish this? Obviously, we could backup everything and completely destroy Ins2 and recreate it on Node2 then rejoin the cluster. But I'm looking for something less destructive.
Is it possible to reinstall SQL, then rejoin the cluster, and then fail Node2 over? Or will there be registry conflictions?
Any help would be appreciated. Also, if you have any links to some official documentation, that would be great too.
Need to set up a linked server in two node SQL Cluster pointing to a standalone SQL thru security setting "be made using current login's security context". It's double hop Kerberos setup. Cluster uses a domain account, so we manually set SPNs for that account with both instance name and protocol as well (set up these SPNs with SQL virtual name only).
Also, constrained delegation has been set to that stand alone SQL services (MSSQLSvc).
Both SQL cluster and standalone servers are in same domain, cluster service account is domain account and standalone SQL uses default SQL local service account.
Coming to SQL, when I create linked server, sometimes it lets me create without issues and sometime it throws this below warning and even if I create it won't work.
Login failed for user 'NT AUTHORITY/ANANOMOUS LOGIN'.
I am new to SQL Server 2012 clustering.I added a new instance to one of the two nodes.when I try to move it to the other node it fails.Do I need to install it on both?If so, what options do you install on the second node?
Is it possible to setup an cluster with sql server 2005 as single node cluster - and lets say in 4 month we add the second node to the cluster? - its because of budget and we do not want to setup then again.
