Month: December 2015

Creating Easy Tier (Multi-tier) Pool with IBM Storwize

Creating an Easy Tier (aka Multi-tier) pool within IBM’s Storwize is pretty simple, just it cannot be done via its GUI. The GUI definitely lacks this functionality, and thus I had to resort to learning some IBM CLI for the Storwize(V5000). I have been told (from IBM) the command(s) are the same for both the Storwize v7000 and v3700 series as well.

The benefits of IBM’s Easy Tier is rather impressive, and I am sure (please correct me if I am wrong) this exists within other SAN vendors as well. In my pool, there are three types of disk drives, SSD and SAS (both enterprise grade and nearline). The benefits of the multi-tiered (easy tier) pool allows data to be (seamlessly) migrated to higher-IO drives/pools that provide higher performance, ie. SSD pools.

In my case this is great, as I will have heavy-hitting IOPS SQL virtual machines that will probably require the higher performing SAS if not SSD drives. And whereas low-IOPS hitting data such as Quorums on nearline drives.

As IBM states, “Easy Tier can automatically migrate data at the sub-LUN/sub-volume level to the most appropriate storage tier. This includes the ability to automatically and non-disruptively relocate logical volume extents with high activity to storage media with higher performance characteristics, while extents with low activity are migrated to storage media with lower performance characteristics.

So how is this all done?

I am going to assume you have already created your mdisk/RAID groups. In my case my SSD drives are a RAID-5 likewise with my SAS (enterprise) drives. RAID-6 for my SAS nearline drives.

Within the CLI, you will need to get the IDs of all your drives, you can do this by running the following command, “lsdrive“. Now you can see all your drives, and their disk types, IDs, etc.

1

Now you need to create your Easy Tier pool with the mkarray command. (More syntax info can be found HERE)

“mkarray -level <raidType> -drive <IDrangeOfDrivesPerDiskTypeGroupSepeartedByColon(s)> <YourPoolName>”

mkarray -level raid5 -drive 0:1:2:3 EasyTier-Pool
*Do not forget to leave 1 drive behind as a spare within your MDisk pool, otherwise you will have no hot-spare, and will have to rebuild*
4 5

Once you have created your pool, and assoicated all the drives to the Easy Tier pool, you can now see all the mdiskX groups, using the lsmdisk command.

2

For more details, use the lsmdiskgrp command.

3

We can now also confirm the Easy Tier pool within the GUI.

image004

Now we can start building/migrating or whatever it is your SAN was designed for! 🙂

 

For more literature on IBM’s Easy Tier, please visit the LINK.

Advertisements

Hyper-V Network Virtual Switches

So you’ve spun up a Windows 2012R2 machine with Hyper-V installed and ready to go. However, now you’re stuck and not sure which type of  Network Virtual Switch (vSwitch) applies to your environment(s)…

In Windows 2012R2, Hyper-V’s network virtual switch runs at Layer 2 (Data Link layer). If you are unfamiliar with this, or either terms, I suggest good old Wikipedia. 🙂 Layer 2 maintains a MAC address table contains the MAC addresses of all the virtual machines (VMs) connected to it. The switch determines where to direct/redirect the packets to based on MAC addresses. It should be noted, in Hyper-V, you can have an unlimited amount of VMs connected to this vSwitch.

In Hyper-V you have three types of Network Virtual Switches: External, Internal and Private. All have similar functions but are disgustingly different.

  1. External vSwitch allows communication between the VMs running within the Hyper-V hosts, the Hyper-V parent partition, and between all VMs on the remote host server. The External vSwitch does require a network adapter on the host (that is not mapped to any other Hyper-V External vSwitch). You can also tag to a VLAN ID.
  2. Internal vSwitch allows communication between all VMs that are connected to the vSwitch and also allows communication between the Hyper-V parent partition. You can also tag to a VLAN ID.
  3. Private vSwitch allows communication between all VMs that are connected to the vSwitch, and that is it. (Note, no communication between the VMs and its Hyper-V parent partition. Also no VLAN ID tagging can occur on the vSwitch)

Without the use of SCVMM (System Center Virtual Machine Manager), I have found there are two ways to go about creating a vSwitch, one via Hyper-V GUI and second via PowerShell.

Let’s start with the GUI:

Launch the Hyper-V console, and right-click on the Hypervisor’s Virtual Switch Manager. Now selecting New virtual network switch, you can specify your properties here. Name your vSwitch, associate to the correct vNIC, tag to the appropriate VLAN ID, etc.

1 vSwitch HyperV Host

You can now specify which vSwitch for your guest VM to use. Within the VMs properties, you will have the option to chose within the Virtual Switch (you will need to create a Network Adapter if not already done). Once selected you can specify your VLAN ID here. (I am finding you cannot specify the VLAN within the Management vSwitch, but it must be done on the client VM’s end) *Again, this is without the use of SCVMM..yet*

2 vSwitch client OS

 

The same process above can be automated via PowerShell. If you’re like me and need to provision a few dozen Hyper-V hosts, creating vSwitches via the GUI is rather tedious. This can be automated with PowerShell (and SCVMM). Please see the code below:

First you will need to get a list of all the Network Adapters your Hyper-V host has to offer. Hopefully you have named them, if you have not, I highly suggest doing this, and considering this best practice and keeping your sanity.

3 Get Adapter names via PS

Once you have the list of vNICs and their names, you can go ahead and start creating vSwitches.

4 Create vSwitch via PS Code 5 Output Create vSwitch via PS

If the code below worked (note only Line 6 is needed to create the External vSwitch) your Hyper-V host should have the vSwitch, or something similar:

1 vSwitch HyperV Host

 

(more…)

OMS Alerting is Here!

Just a few days ago, Microsoft introduced OMS alerting. For people unaware, this is native alerting for Operations Management Suite. This feature has been requested since OMS’ inception, and Microsoft has listened and finally delivered! Check out their blogpost HERE.

I will be setting up some labs and testing this out in the next few days.

Load Balancing SCOM Agents

So you have multiple SCOM Management Servers, yet you just happen to have all of your SCOM agents reporting to one server, or maybe two if you half tried to load balance your agents. There are several reasons why you would want to have multiple Management Servers, ie. off-load workflows, reduce stress on servers, etc., etc. Well what is the point of having multiple Management Servers yet nearly all of your agents are reporting to one, or maybe two at best Management Servers, while the others are collecting dust. Load balance those agents! You could manually move an agent by right clicking and moving to a new server, or you could let our friend PowerShell automate this for you.

In my experience I have seen many SCOM environments where load balancing is either done manually, or not done at all. And usually manually implies the SCOM administrator takes a look which of the servers has the least agents, and deploys away. That works, but why not deploy to any server then let PowerShell load balance for you.

In the solution below, I am using PowerShell along with Orchestrator 2012R2. The runbook can be setup to run ad-hoc, or run regularly, ie. monthly, weekly, etc. Of course if you do not Orchestrator deployed in your environment, you could very well take the script and schedule it to run via Windows Scheduled tasks.

ce63742c-85d7-402e-b114-c3979b7ce32b

Here I have created a Runbook to execute the script, and then send back a warning notification if the Runbook failed, or an informational notification that the Runbook executed successfully.

See below for the PowerShell script. Please note, you will need to change the Line 5 with a SCOM Management server applicable to your environment, duh. This script can also be modified, and you can load balance between two gateway servers.

The script can be found HERE!

Happy SCOM’ing!