 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
SQL Server Forum / Other Technologies / Clustering / July 2005Moving Quorum Drive
Hi, I have a two node failover cluster with two external disk arrays. I have these two array for some special testing. I no longer need both of them and need to do some specific testing with just one array. My problem is that the array I need to test with has all my SQL data but the other array has the quorum drive. What is the best method to moving the quorum drive to the other array? Also, is it possible to have active/active SQL cluster where both SQL servers (cluster nodes) are sharing the same SQL data files? Thank you in advance. Rodney or one of the other Cluster MVPs will have to anwer the quorum disk move question. The Active/Active nomenclature is misleading and obsolete. There is no sharing of data files between SQL instances within a cluster (or anywhere else for that matter). Geoff N. Hiten Microsoft SQL Server MVP > Hi, > [quoted text clipped - 9 lines] > > Thank you in advance. So does that mean if you need to manage extremely high availability with 900K+ users using SQL, the only option you have is to get beefier boxes? and live with the 40 second downtime you will get from the cluster failover? Please note that most of these users are not humans and they know how to hammer a box :-) Thanks. > Rodney or one of the other Cluster MVPs will have to anwer the quorum disk > move question. [quoted text clipped - 19 lines] >> >> Thank you in advance. SQL is a scale up not a scale out technology. SQL 2005 has some new tricks to give you more high availability options, but you are still looking at buying larger boxes as you outgrow your existing ones. I wouldn't worry too much, the new TCP-C benchmark released last month by Microsoft is several times the average daily transaction volume of Visa and it was run on a single system. Besides, after a certain scale, the disk subsystem and license costs far outweigh the host computer costs on almost any system, regardless of which DMBS you are using. Geoff N. Hiten Microsoft SQL Server MVP. > So does that mean if you need to manage extremely high availability with > 900K+ users using SQL, the only option you have is to get beefier boxes? [quoted text clipped - 27 lines] >>> >>> Thank you in advance. Thank you Geof. Would you happen to have a link to the benchmark study? > SQL is a scale up not a scale out technology. SQL 2005 has some new > tricks to give you more high availability options, but you are still [quoted text clipped - 39 lines] >>>> >>>> Thank you in advance. http://www.tpc.org/tpcc/results/tpcc_result_detail.asp?id=105060604 GNH > Thank you Geof. Would you happen to have a link to the benchmark study? > [quoted text clipped - 41 lines] >>>>> >>>>> Thank you in advance. > So does that mean if you need to manage extremely high availability with > 900K+ users using SQL, the only option you have is to get beefier boxes? > and live with the 40 second downtime you will get from the cluster > failover? Please note that most of these users are not humans and they > know how to hammer a box :-) You can do both. You can scale up, and you can scale out by using multiple instances of SQL. However, to scale out, you must be able to break up databases between the instances. If everyone requires access to the same database, then you don't have the ability to scale out. 900K+ users? I am only guessing that there are not that many simultaneous users. How many queries a second are we talking about, just out of curiosity? Thank you Russ. Regarding 900K users, unlike humans, these are very SQL chatty clients and can request to run several SPs at once. This is why I am planning on creating a hierarchy and breaking the databases into multiple instances (just as you suggested). I will probably use concentrator SQL boxes so I am not killing a single box. :-) This way my whole environment is not dependent on a single point of failure (SQL Server). >> So does that mean if you need to manage extremely high availability with >> 900K+ users using SQL, the only option you have is to get beefier boxes? [quoted text clipped - 10 lines] > users. How many queries a second are we talking about, just out of > curiosity? I believe Russ answered it already, sorry I have been on vacation :) Cheers, Rod MVP - Windows Server - Clustering http://www.nw-america.com - Clustering http://www.msmvps.com/clustering - Blog > Rodney or one of the other Cluster MVPs will have to anwer the quorum disk > move question. [quoted text clipped - 19 lines] >> >> Thank you in advance. can't help your sql question, but hopefully this will help you to move your quorum drive http://support.microsoft.com/default.aspx?scid=kb;en-us;280353 > Hi, > [quoted text clipped - 9 lines] > > Thank you in advance. Thank you Seth. > can't help your sql question, but hopefully this will help you to move > your quorum drive [quoted text clipped - 14 lines] >> >> Thank you in advance. > Hi, > [quoted text clipped - 4 lines] > has the quorum drive. What is the best method to moving the quorum drive > to the other array? The best method of moving the quorum, in my opinion, is to simply use the GUI. Using the GUI requires that both arrays be visible to the nodes and that the physical disk that you want to use be also visible to both nodes. In cluster admin, right click the cluster name, select the Quorum tab and use the drop down box to select what drive you want to host the quorum. > Also, is it possible to have active/active SQL cluster where both SQL > servers (cluster nodes) are sharing the same SQL data files? Active/active? No. It isn't possible for two instances to share the same databases and logs. Thank you Russ. I will defo give it a try as soon as I create some space on my DB array. :-) It sound way to simple. hehe >> Hi, >> [quoted text clipped - 16 lines] > Active/active? No. It isn't possible for two instances to share the same > databases and logs. Hi Russ. I am trying to follow what you recommend and came across an issue. Could you please see if you can shed some light on this? If anyone else would like to chime in, you are more than welcome. I am attaching a few screenshots to help a little. Essentially I am unable to add a new disk resource (P:). I would like to use this resource as my quorum drive. here is what I have done so far. My Boot drive is D:. Current Quorun is on Q: that is a partition on HP Array. S: Drive is a data Drive on the same HP Array. E: is my data drive on my EMC Array. This is the drive that was using whole SCSI disk as the drive. I copied my data out, deleted the partition and recreated it as a smaller drive so that I have extra space for the new Quorum. Created P drive as the new Quorum. I am trying to get rid of HP Array (Q and S drives). When I try to create a new resource, I do not see my newly created P drive in there at all. I tried to move the group over to my second server and I can see the P drive from both servers but not when adding a new resource. Thank you. > Thank you Russ. > [quoted text clipped - 23 lines] >> Active/active? No. It isn't possible for two instances to share the same >> databases and logs. Sorry, it wouldn't let me send the post with attachment so here are the link to screenshots. http://web1.greatbasin.net/~kayas/SharesRes.JPG http://web1.greatbasin.net/~kayas/EMC.JPG http://web1.greatbasin.net/~kayas/dskmgr.JPG Thank you. > Hi Russ. > [quoted text clipped - 56 lines] >>> Active/active? No. It isn't possible for two instances to share the same >>> databases and logs. Microsoft clustering does not like partitions. P is a partition. You will only see E as a disk resource. You will never be able to add P as a disk resource, since it is already configured as E: Next issue, to move the Quorum from Q to P, E Disk Resource would have be reside in the Shared Array Cluster S... Group. You could add a drive or LUN to shared array and make it P and move the Quorum to it. Cheers, Rod MVP - Windows Server - Clustering http://www.nw-america.com - Clustering Website http://msmvps.com/clustering - Blog > Hi Russ. > [quoted text clipped - 56 lines] >>> Active/active? No. It isn't possible for two instances to share the same >>> databases and logs. Got it. Thanks. Since I do not have any additional disks, I will probably delete and recreate E drive with a smaller size at the array level. This should give me enough space to create a New Quorum and show up in Disk Manager as a new physical drive. Do you think I can simply take E drive offline and the delete/recreate it without doing anything else in the cluster Admin? thanks. > Microsoft clustering does not like partitions. P is a partition. You will > only see E as a disk resource. You will never be able to add P as a disk [quoted text clipped - 78 lines] >>>> same >>>> databases and logs. By looking at your graphics you have SQL installed and running on your E: drive, so I would say you would indeed hurt things if you delete it. Uninstall SQL or move it too. Cheers, Rod MVP - Windows Server - Clustering http://www.nw-america.com - Clustering Website http://msmvps.com/clustering - Blog > Got it. Thanks. > [quoted text clipped - 93 lines] >>>>> same >>>>> databases and logs. I have stopped SQL through Cluster, and copied out everything onto another location. I did this when I reformatted E to create P. What I am concerned with is that, will Cluster look at the drive I will create just as it does now? I will recreate it with the same drive letters. > By looking at your graphics you have SQL installed and running on your E: > drive, so I would say you would indeed hurt things if you delete it. [quoted text clipped - 107 lines] >>>>>> same >>>>>> databases and logs. For SQL only: If you create a new LUN and map it to the old letter and set the dependencies correctly and copy the data exactly, SQL will think it is the same drive and will work correctly. Note that this works with SQL only and may not work with any other clustered resource. Also note that there are a lot of "IF"s in that statement. Geoff N. Hiten Microsoft SQL Server MVP >I have stopped SQL through Cluster, and copied out everything onto another >location. I did this when I reformatted E to create P. What I am concerned [quoted text clipped - 114 lines] >>>>>>> same >>>>>>> databases and logs. > For SQL only: If you create a new LUN and map it to the old letter and > set the dependencies correctly and copy the data exactly, SQL will think [quoted text clipped - 4 lines] > Geoff N. Hiten > Microsoft SQL Server MVP It also works with Exchange. AWESOME............... I deleted the LUN and recreated multiple ones. Created new drives, and shared data drive with the same drive letter. After a reboot of both nodes everything looks just peachy. Simply moved the quorum to one of the new LUN/Drives and now I can shut off the other Disk array and still have my cluster up and running. :-) Thank you everyone for all your help. :-) >> For SQL only: If you create a new LUN and map it to the old letter and >> set the dependencies correctly and copy the data exactly, SQL will think [quoted text clipped - 6 lines] > > It also works with Exchange. |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2009 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.