Kenny's questions

There appears to be a great deal of conflicting information about whether or not SAN is required for a Sql Server 2012 Mirroring setup.

I currently have an Active-Passive-Witness SQL 2008R2 setup where each of the machines in the setup have plain old DAS raid arrays.

I would like to implement exactly the same again with SQL Server 2012, however much of the online documentation (e.g. Prerequisites, Restrictions, and Recommendations for AlwaysOn Availability Groups (SQL Server) - http://msdn.microsoft.com/en-us/library/ff878487.aspx) suggests each node in the "AlwaysOn Availability Group" must be a node of a Windows Server Failover Clustering (WSFC) cluster.

I'm not trying to set up a SQL Server Failover Cluster - only database mirroring between two instances of SQL Server (active-passive-witness).

When I look up the requirements for setting up a Windows Server Failover Clustering (WSFC) cluster I always see SAN / shared storage as a requirement.

However I have seen a pretty definitive answer to my question here: "AlwaysOn Availability Groups do not require deployment of a Failover Cluster Instance or use of symmetric shared storage (SAN or SMB)." - http://msdn.microsoft.com/en-us/library/hh270278.aspx - but am unsure what to believe.

Has anybody out there tried an Active-Passive-Witness mirroring setup with SQL Server 2012 and "AlwaysOn Availability Groups" where you didn't need to use SAN / shared storage?

Many thanks,

Kenny

| tape drive 1 | ---- | switch 1 | ---- hba1 | backup server | hba2 ---- | switch 2 | -----| tape drive 2 |

Our backup server has 2 FC HBAs. Each is zoned into a different drive on the back of the tape library via a separate switch.

Each tape is connected to a different switch presumably to allow for the switch to fail yet allow the backup server to still have one accessible tape.

We seem to have lost an HBA. As a temporary measure, is it possible to move the tape drive into the other switch which is still connected to the backup server and zone the one HBA into both FC Tape drives?

---- | switch 1 | ---- hba1 | backup server | hba2 ---- | switch 2 | -----| tape drive 2 | + | tape drive 1 |

Will this confuse Backup Exec seeing both tape drives coming from one HBA, or dos EMC's HBAnywere software or drivers abstract this away knowledge away from the application and just present the tape drives to the Operating System regardless of which HBA they are appearing..?

My suspicion would be that you can have as many LUNS as you want (within OS Max) coming from one HBA or the other HBA and probably Backup Exec wouldn't know the difference, but I can't risk just changing these things without the concrete knowledge that I'm not going to make things worse. I don't want to risk Backup Exec's access to the last working tape drive.

Thanks!

I'm trying to plan a SQL Server 2008 R2 2 node Active/Passive failover cluster.

Is it possible to set up multiple named instances running on the cluster? e.g.

sqlcluster
sqlcluster\demo
sqlcluster\dev

And could demo and dev share the same SAN LUN?

Thanks,

Kenny

I want to change SSL bindings to a replacement certificate, but don't want to affect logged in users.

Will doing this cause the Application Pools for the web site to recycle?

Many thanks,

Kenny

If I make a change to the file:

%windir%\system32\inetsrv\config\applicationHost.config

and update it with a change resulting in a valid config, will it cause IIS7 to restart and force restarts on all the apps?

I am trying this out with a hot-cloned vm of the server, but like to know if anyone has experience of either of these scenarios, what the expected outcome should be.

Many thanks

Mouse is pointing to a fraction of where it should in a VMware vSphere 4.0 vm with the windows vSphere client.

Booting up UBCD -> PartedMagic... The mouse pointer has x/y co-ordinates of a smaller fraction of that which it "should" have.

This can be seen here: http://www.youtube.com/watch?v=Bx2emQ1IpQw

Can anyone work out how I can make the mouse work properly with GParted and vSphere 4?

I know I could probably spend ages learning how to use the command line to do what I need here, but the whole point of using GParted is for the GUI..

Cheers!

Do you know where to download EMC's Switch Analysis Tool (SWAT)?

I've followed the instructions in the "Customer Advisory Letter" I have regarding FLARE upgrade, but the menu options in EMC PowerLink are not there.

The letter states I should find SWAT within Powerlink in "Home > Support > Product and Diagnostic Tools > Environment Analysis Tools > Switch Analysis Tool (SWAT). " - but this menu structure no longer exists, and I've tried searching the site, to no avail.

https://community.emc.com/thread/113532

Many thanks,

Kenny

I have discovered a Description entry ("EMC Flex 8-16") in an systems audit which I can't identify. I have asked the Vendors, but have not yet heard back. I'll post an answer myself when I find out.

I have been unable to find info via searching the web (or EMC's support site).

So for myself and others who may in the future attempt to identify the same thing, does anyone know what it is?

I have a "gut feeling", but would love to hear from anyone who actually knows.

Many thanks,

Kenny

Summary:

My "dual-core" VM has one core being totally wasted by a constantly running Dr Watson process (provided by the .Net Runtime), and I can't figure out how to stop if from running.

When it runs in conjunction with the timed sharepoint search service (once every 10min), they both consume 2 cores worth of CPU causing denial of service to the IIS processes that service the users of the box, therefore making Sharepoint entirely unusuable.

Detail:

dw20.exe starts up all the time, and it runs constantly at 100% of one core.

• Dedicated Sharepoint (web front end and index server) VMware VM.
• Windows Server 2008,
• Msinfo32 output: http://pastebin.org/410798

I have the identical problem to this fellow: How do you shut Dr. Watson reporting off on Windows Server 2008?, but following the link provided and turning off Windows Error Reporting does not stop it from running.

The process is an executable called dw20.exe, and it lives in: C:\Windows\Microsoft.NET\Framework64\v2.0.50727

I have tried renaming or deleting this file, via several methods (logged in as Admin), but cannot seem to do it. Save taking the box down, and mounting the filesystem under linux and renaming it, I can't think now of a way to stop it running. I've tried Unlocker 1.8.9, Killbox, command line (right-click, run as Administrator).

Can you think of how I can stop this process DOSing my Sharepoint box, keeping the current architecture? (I realise by setting up another Sharepoint machine to do the indexing, this would stop the index service using a core...)

I suppose I could try renaming the executable by booting the server into the recovery console, or mounting the windows filesystem under a linux livecd, but then I'd worry about the process that trying to execute it in the first place - hopefully that won't go crazy when it can't find it...

On a Dell SAS 5/ir RAID controller, I have a RAID 1 array which is Degraded.

One drive has failed.

Upon replacing this drive the controller doesn't automatically resync the array.

There's an option to manually add the new drive to the array, but when I try this, the controller bios shows:

Screenshots: http://imagebin.ca/view/QdCuJI5.html

WARNING! Data on physical disks that are part of the virtual disk will be LOST! Are you sure you want to commit RAID Configuration changes?

Now - it is obvious that the drive I put in to replace the failed one will have data loss, as it should get data from the working drive synced onto it.

However, the Controller's message doesn't state what drive there'll be data loss. It could be that making this change will create a fresh RAID LUN with no data. I don't know, and the controller doesn't make it clear.

Can anyone say for definite that this will not affect the data on the original drive?

It just worries me that the message states "physical disks" - plural, and not "Data on the secondary physical disk that you just added will be LOST!"

No doubt this is what the message means, but I don't like guessing with critical data!

Many thanks!

A drive failed on a Dell SAS 5i controller - see controller bios screengrab: http://imagebin.ca/view/SkZbszA.html

The primary is a 160GB 10k sata drive.

I added a 250GB 7k rpm drive in the hope that the array would rebuild onto this drive. This did not happen. (assuming that the controller would just operate at the speed of the slowest drive)

The controller could see the new drive, but it didn't automatically rebuild the raid1 onto this drive. (my assumption is that it did not do this rebuild as the drive sizes are different).

There was an option to add the new drive to the existing raid1 array - but when I tried this a message appeared stating that all data on the array would be lost. (I didn't get a screenshot of this message, I will later)

Should the SAS 5i allow me to rebuild the array onto a larger drive?

Is the option to add the drive to the array the right way to go?

Many thanks!

I have a Brocade SAN Fabric Switch, a 200E.

When using the web based management interface "SwitchExplorer" I can click the port, and I see "Port Administration Services".

In the first screen of data that appreas, there's a row called "Health" which has value "Unmonitored".

Do you know how to set this port to be Monitored? And if also - what "Health" monitoring does?

I'm hoping it'll email or log if there's a connection or disconnection..

Many thanks for looking...

Kenny

I'm using Active Directory to log into RHEL. To do this, I followed the steps outlined here:

http://www.markwilson.co.uk/blog/2007/05/using-active-directory-to-authenticate-users-on-a-linux-computer.htm

I'd like to be able to read data from Windows Servers shared folders without being prompted for a password.

On Windows I log into an AD domain, and when I access windows file shares on a server on the LAN (also part of the AD domain) my I can just access them with no authentication step.

I've used SMBclient on Linux to access these shares, but it asks for my password.

I would like to be able to script access to the data on the shares, but I can't if there's a password prompt in the way. Well, I could, but it's not how I want to do it.

Now, since I'm logged in using my active directory username & password, can't I just access the shares without jumping that extra hoop?

I know I can mount the share using something like:

//192.168.0.5/share /mnt/windows cifs auto,username=steve,password=secret,rw 0 0

but access will depend who is logged in... each user logging in should have their own unique AD access privelages.

Thanks for reading!

Do you know how to set a free disk (big enough to be a member of all RAID groups) to be a "Hot Spare" available to all RAID groups in a Clariion CX3-20 using Navisphere 6?

I have created a RAID group with number of drives "1" and type "Hot Spare", but get the feeling that there's something else I need to do, to allow it to be used automatically by all the RAID groups in my Clariion.

I have read that one must make sure SATA disks don't become Hot Spares for FC RAID groups, and vica versa, and although my Array is all FC, I get the feeling that to set that up, there must be some notion of telling the Clariion which RAID groups (of type Hot Spare) can be Hot Spares for with other RAID groups.

Many thanks for reading!