Tag Archives: performance

PART 1 – Problems with RAID and Object Based Storage for data protection

Posted on by
2

I regularly get asked to compare the resiliency of traditional centralized storage with converged as well as newer technologies such as hyper-converged.

So this post will discuss the problems with RAID and newer hyper-converged solutions using Object based storage for data protection.

This post will discuss two examples below, with Part 2 discussing Hyper-converged solutions using Distributed File Systems.

1. Traditional RAID

2. Hyper-converged Object Based Storage

Starting with Traditional shared storage, and the most common RAID level in my experience, RAID 5.

The below diagram shows a 3 x 4TB SATA drives in a RAID 5 with a Hot Spare.
3 Disk R5 w Hot Spare NO BG

Now lets look a drive failure scenario. We now have the Hot Spare activate and start rebuilding as shown below.

3 Disk R5 w Hot Spare REBUILDING NO BG

So this all sounds fine, we’ve had a drive failure, and a spare drive has automatically taken its place and started rebuilding the data.

The problem now is that even in this simplified/small example we have 2 drives (or say 200 IOPS of drives) trying to rebuild onto just a single drive. So the maximum rate at which the RAID 5 can restore resiliency is limited to that of a single drive or 100 IOPS.

If this was a 8 disk RAID 5, we would have 7 drives (or 700 IOPS) trying to rebuild again to only a single drive or 100 IOPS.

There are multiple issues with this architecture.

  1. The restoration of resiliency of the entire RAID is constrained by the destination drive, in this case a SATA drive which can sustain less than 100 IOPS
  2. A single subsequent HDD failure within the RAID will cause data loss.
  3. The RAID rebuild is a high impact activity on the storage controllers which can impact all storage
  4. The RAID rebuild is an especially high impact activity on the virtual machines running on the RAID.
  5. The larger the RAID or the capacity drives in the RAID, the longer the rebuild takes and the higher the performance impact and chance of subsequent failures leading to data loss.

Now I’m sure most of you understand this concept, and have felt the pain of a RAID rebuild taking many hours or even days, but with new hyper converged technology this issue is no longer a problem, right?

Wrong!

It entirely depends on how data is recovered in the event of a drive failure. Lets look at an example of an hyper-converged solution using an object store.The below shows a simplified example of a Hyper-converged Object Based Storage with 4 objects represented by Object A,B,C and D in Black, and the 2nd replicated copy of the object represented Object A,B,C and D in Purple.

Note: Each object in the Object Store can be hundreds of GB in size.HyperconvergedObjectStoreNormal

Let’s take a look what happens in a disk failure scenario.

HyperconvergedObjectStoreFailure

From the above diagram we can see a drive has failed on Node 1, which means Object A and Object D’s replica have been lost. The object store will then replicate a copy of Object A to Node 4, and a replica of Object D to Node 2 to restore resiliency.

There are multiple issues with this architecture.

  1. Object based storage can lack granularity as Objects can be 200Gb+.
  2. The restoration of resiliency of any single object is constrained by the source drive or node.
  3. The restoration of resiliency of any single object is also constrained by the destination drive or node.
  4. The restoration of multiple objects (such as Object A & D in the above example) is constrained by the same drive or node which will result in contention and slow the process of restoring resiliency to both objects.
  5. The impact of the recovery is High on virtual machines running on the source and destination nodes.
  6. The recovery of an Object is constrained by the source and destination node per object.
  7. Object stores generally require a witness, which is stored on another node in the cluster. (Not illustrated above)

It should be pointed out, where SSDs are used for a write cache, this can help reduce the impact and speed up recovery in some cases, but where data needs to be recovered from outside of cache, i.e.: A SAS or SATA drive, the fact writes go to SSD makes no difference as the writes are constrained by the read performance.

Summary:

Traditional RAID used by SAN/NAS and newer Hyper-converged Object based storage both suffer similar issue when recovering from drive or node failures which include:

  1. The restoration of resiliency is constrained by the source drive or node
  2. The restoration of resiliency is constrained by the destination drive or node
  3. The restoration is high impact on the desination
  4. The recovery of one object is constrained by the network connectivity between just two nodes.
  5. The impact of the recovery is High on any data (such as virtual machines) running on the RAID or source/destination node/s
  6. The recovery of RAID or an Object is constrained by a single part of the infrastructure being a RAID controller / drive or a single node.

In Part 2, we will look at the Hyper-converged Distributed File Systems.

Rule of Thumb: Sizing for Storage Performance in the new world.

Posted on by
6

In the new world where storage performance is decoupled with capacity with new read/write caching and Hyper-Converged solutions, I always get asked:

How do I size the caching or Hyper-Converged solution to ensure I get the storage performance I need.

Obviously I work for Nutanix, so this question comes from prospective or existing Nutanix customers, but its also relevant to other products in the market, such as PernixData or any Hybrid (SSD+SAS/SATA) solution.

So for indicative sizing (i.e.: Presales) where definitive information is not available and/or where you cannot conduct a detailed assessment , I use the following simple Rule of Thumb.

Take your last two monthly full backups, and take the delta between them and multiply that by 3.

So if my full backup from August was 10TB and my full backups from September is 11TB, my delta is 1TB. I then multiply that by 3 and we get 3TB which is our assumption of the “Active Working Set” or in basic terms, the data which needs performance. (Because cold or inactive data can sit on any tier without causing performance issues).

Now I  size my SSD tier for 3TB of usable capacity.

The next question is:

Why multiple the backup data delta by 3?

This is based on an assumption (since we don’t have any hard data to go on) that the Read/Write ratio is 70% Read, 30% write.

Now those of you familiar with this thing called Maths, would argue 70/30 is 2.33333 which is true. So rounding up to 3 is essentially a buffer.

I have found this rule of thumb works very well, and customers I have worked with have effectively had All Flash Array performance because the “Active Working Set” all resides within the SSD tier.

Caveats to this rule of thumb.

1. If a customer does a significant amount of deletions during the month, the delta may be smaller and result in an undersized SSD tier.

Mitigation: Review several months of full backup logs and average the delta.

2. If the environment’s Read/Write ratio is much higher than 70/30, then the delta from the backup multiplied by 3 may again result in  an undersized SSD tier.

Mitigation: Perform some investigation into your most critical workloads and validate or correct the assumption of multiplying by 3

3. This rule of thumb is for Server workloads, not VDI.

VDI Read/Write ratio is generally almost opposite to server, and around 30/70 Read/Write. However the SSD tier for VDI should be sized taking into account the benefits of VAAI/VCAI cloning and things like de duplication (for Memory and SSD tiers) which some products, like Nutanix offer.

Summary / Disclaimer

This rule of thumb works for me 90% of the time when designing Nutanix solutions, but your results may vary depending on the platform you use.

I welcome any feedback or suggestions of alternate sizing strategies which I will update the post with where appropriate.

Is VAAI beneficial with Virtual Storage Appliance (VSA) based solutions ?

Posted on by
7

I saw a tweet recently (below) which inspired me to write this post as there is still a clear misunderstanding of the benefits VAAI provides (even with Virtual Storage Appliances).

vaaionvsatweet2

I have removed the identity of the individual who wrote the tweet and the people who retweeted this as the goal of this post is solely to correct what I believe is mis-information.

My interpretation of the tweet was (and remains) if a solution uses a Virtual Storage Appliance (VSA) which resides on the ESXi host then VAAI is not providing any benefits.

My opinion on this topic is:

Compared to a traditional centralised NAS (such as a Netapp or EMC Isilon) providing NFS storage with VAAI-NAS support, a Nutanix or VSA solution has exactly the same benefits from VAAI!

My 1st reply to the tweet was:

vaaionvsatweetconvojoshreply2

The test I was referring to with Netapp OnTap Edge can be found here which was posted in Jan 2013, well prior to my joining Nutanix when I was working for IBM where I had been evangelising VAAI/VCAI based solutions for a long time as VAAI/VCAI provides significant value to VMware customers.

The following shows the persons initial reply to my tweet.

vaaionvsatweetconvo32

I responded with the below mentioning I will do a blog which is what you’re reading now.

I went onto provide some brief replies as shown below.

repliesdetail

The main comments from this persons tweets I would summarize (rightly or wrongly) below:

  • VAAI is designed only to offload functions externally (or off the ESXi host)
  • He/She had not seen any proof of performance advantages from VAAI on VSAs
  • Its broken logic to use VAAI with a VSA

Firstly, I would like comment on VAAI being designed to offload functions externally (or off the ESXi host). I don’t disagree VAAI has some functions designed to offload to the (centralised) array but VAAI also has numerous functions which are designed to bring other efficiencies to a vSphere environment.

An example of a feature designed to offload to a central array is the “XCOPY” primitive.

A simple example of what “XCOPY” or Extended Copy provides is offloading a Storage vMotion on block based storage (i.e.: VMFS over iSCSI,FC,FCoE not NFS) to the array so the ESXi host does not have to process the data movement.

This VAAI primitive would likely be of little benefit in a VSA environment where the storage is presented is block based and Storage DRS for example was used. The data movement would be offloaded from ESXi to the VSA running on ESXi and host would still be burdened with the SvMotion.

However XCOPY is only one of the many primitives of VAAI, and VAAI does alot more than just offload Storage vMotions.

For the purpose of this post, I will be discussing VAAI with Nutanix whos Software defined storage solution runs in a VM on every ESXi host in a Nutanix cluster.
Note: This information is also relevant to other VSAs which support VAAI-NAS.

So what benefit does VAAI provide to Nutanix or a VSA solution running NFS?

Nutanix deploys by default with NFS and supports the VAAI-NAS primitives which are:

  • Full File Clone
  • Fast File Clone
  • Reserve Space
  • Extended Statistics

Note: XCOPY is not supported on NFS, importantly and specifically speaking for Nutanix it is not required as SvMotion will be rarely if ever used with Nutanix solutions.

See my post “Storage DRS and Nutanix – To use, or not to use, that is the question?” for more details on why SvMotion is rarely needed when using Nutanix.

For more details of VAAI primitives, Cormac Hogan (@CormacJHogan) wrote an excellent post which can be found here.

Now here is an example of a significant performance benefits of VAAI with Nutanix.

Lets look at Clone of a VM on a Nutanix platform, the VMs details are below.admin01vm

The VM I have used for this test resides on a datastore called “Management” (as per the above image) which presented via NFS and has VAAI (Hardware Acceleration) enabled as shown below.datastore

Now if I do a simple clone of a VM (as shown below) if the VM is turned on, VAAI-NAS is bypassed as the “Fast File Clone” primitive only works on VMs which are powered off.

clone

So a simple way to test the performance benefits of VAAI on any platform (including Hyper-converged such as Nutanix, a Virtual Storage Appliance (VSA) such as Netapp Ontap Edge or traditional centralised SAN or NAS) is to clone a VM while powered on then shut-down the VM and clone it again.

I performed this test and the first clone with the VM powered on started at 1:17:23 PM and finished at 1:26:12 PM, so a total of 8 mins 49 seconds.

Next I shut down the VM and repeated the clone operation.cloneresults

As we can see in the above screen capture from the 2nd clone started at 1:26:49 PM and finished at 1:26:54 PM, so a total of 5 seconds.

The reason for the huge difference in the speed of the two clones is because VAAI-NAS “Fast File Clone” primitive offloaded the 2nd clone to the Nutanix platform (which runs as a VM on the ESXi host) which has intelligently cloned the VM (using metadata resulting in almost zero data creation) as opposed to 1st clone where VAAI-NAS was not used which resulted in the hypervisor and storage solution having to read 11.18GB of data (being the source VM – Admin01) and write a full copy of the same data resulting in effectively >22GB of data movement in the environment.

Now from a capacity savings perspective, a simple way to demonstrate the capacity savings of VAAI on any platform is to clone a VM multiple times and compare the before and after datastore statistics.

Before I performed this test I captured a baseline of the Management datastore as shown below.

BeforeCloningCapacityVMcount

The above highlighted areas show:

  • Virtual Machines and Templates as 83
  • Capacity 8.49TB
  • Provisioned Space 7.09TB
  • Free Space 7.01TB

I then cloned the Admin01 VM a total of 7 times.clone7vmsrecenttasks

Immediately following the last clone completing I took the below screen shot of the Management datastores statistics.

AfterCloningCapacityVMcount

The above highlighted areas in the updated datastore summary show:

  • Virtual Machines and Templates INCREASED by 7 to 90 (as I cloned 7 VMs)
  • Capacity remained the same at 8.49TB
  • Provisioned Space INCREASED to 7.29TB as we cloned 7 x ~40Gb VMs (Total of ~280GB)
  • Free Space REMAINED THE SAME at 7.01TB due to VAAI-NAS Fast File Clone primitive working with the Nutanix Distributed File System.

So VAAI-NAS allowed a VM of ~11GB of used storage (~40GB provisioned) to be cloned without using any significant additional disk space and the clones were each done in between 5 and 7 seconds each.

So some of the benefits VAAI-NAS provides to Nutanix (which some people would term as a VSA type solution) include:

  • Near instant VM cloning via vSphere Client/s (as shown above)
  • Near instant Horizon View Linked Clone deployments (VCAI) – Similar to example shown.
  • Near instant vCloud Director clones (via FAST Provisioning) – Similar to example shown.
  • Major capacity savings by using Intelligent cloning rather than Full Clones (As shown above)
  • Lower CPU overhead for both ESXi hosts AND Nutanix Controller VM (CVM)
  • Ability to create EagerZeroThick VMDKs on NFS (e.g.: To support Fault Tolerance & clustered workloads such as Oracle RAC)
  • Enhanced ability to get statistics on file sizes , capacity usage etc on NFS

In Summary:

Overall I would say that VMware have developed an excellent API in VAAI and Nutanix along with VSA providers having support for VAAI provides major advantages and value to our joint customers with VMware.

It would be broken logic NOT to leverage the advantages of VAAI regardless of storage type (VSA, Nutanix or traditional centralized SAN/NAS) and for the vast majority of vSphere deployments, any storage solution not supporting (or having issues/bugs with) VAAI will have significant downsides.

I am looking forward to ongoing developments from VMware such as vVols and VASA 2.0 to continue to enhance storage of vSphere solutions in the future.

I hope customers and architects now have the correct information to make the most effective design and purchasing recommendations to meet/exceed customer requirements.