Integrity of I/O for VMs on NFS Datastores – Part 1 – Emulation of the SCSI Protocol

This is the first of a series of posts covering how the Integrity of I/O is ensured for Virtual Machines when writing to VMDK/s (Virtual SCSI Hard Drives) running on NFS datastores presented via VMware’s ESXi hypervisor as a “Datastore”.

Note: To be crystal clear, this post is not talking about presenting NFS direct to Windows or any other guest operating system.

This process is patented (US7865663) by VMware and its inventors and on the patent the process is called “SCSI Protocol Emulation”.

This series will first cover the topics in a vendor agnostic manner, meaning I am talking in general about VMware + any NFS storage on the VMware HCL with NFS support.

Following the vendor agnostic posts, I will follow with a series of posts focusing specifically on Nutanix, as the motivation for the series was to cover off this topic for existing or potential Nutanix customers, some of whom are less familiar with NFS and have asked for clarification, especially around virtualizing Business Critical Applications (vBCA) such as Microsoft SQL and Exchange.

The below diagram visualizes shows how storage can be presented to an ESXi host and what this series will focus on.

A VM accesses its .vmx and .vmdk file/s via a datastore the same way, regardless of the underlying storage protocol (DAS SCSI, iSCSI , NFS , FCP).

GUID-AD71704F-67E4-4AC2-9C22-10B531755566-high

In the case of NFS datastores, SCSI protocol emulation is used to allow the Guest Operating System (OS) and application/s to read and write via SCSI even when the underlying storage (which is abstracted by the hypervisor) is served via NFS which does not natively support the same commands.

Image Source: https://pubs.vmware.com/vsphere-50/index.jsp?topic=%2Fcom.vmware.vsphere.introduction.doc_50%2FGUID-2E7DB290-2A07-4F54-9199-B68FCB210BBA.html

In the following section, and throughout this series, many images shown are from the patent (US7865663) and are the property of the patent owners, not the author of this article.

The areas which I will be focusing on are the ones where there has been the most concern in the industry, especially for business critical applications, such as Microsoft SQL and Microsoft Exchange, being how are the VM operating system and application/s (or data integrity) are impacted when issuing commands when the storage is abstracted by the hypervisor and served to via NFS which does not have equivalent I/O commands as SCSI.

Some examples areas of concern around the industry for VMs running on datastores backed by NFS are:

1. SCSI Aborts / Resets
2. Forced Unit Access (FUA) & Write Through
3. Write Ordering
4. Torn I/O (Writes + Reads)

In this first part, we will look at the SCSI Protocol Emulation process and discuss SCSI Aborts and Resets and how the SCSI protocol emulation process deals with these.

Below is a diagram showing the flow of an I/O request for a VM writing SCSI commands to a VMDK (formatted as NTFS) through the SCSI emulation process and through to the NFS storage.

US07865663-20110104-D00005

The first few steps in my opinion are fairly self explanatory, where it gets interesting for me, and one of the points of contention among I.T professional (being SCSI aborts) is described in the box labelled “550“.

If the SCSI command is an abort (which has no equivalent in the NFS protocol), the SCSI emulation process removes the corresponding request from the virtual SCSI request list created in the previous step (box labelled “540“).

The same is true if the SCSI command is a reset (which also has no equivalent in the NFS protocol), the SCSI emulation process removes the corresponding request from the virtual SCSI request list. This process is shown below in the box labelled “560

US07865663-20110104-D00006

Next lets look at what happens if the SCSI “abort” or “reset” command is issued after the SCSI emulation process has passed on the command to the storage and is now receiving a reply to a command which the Guest OS / Application has aborted?

Its quite simple, the SCSI emulation process receives a reply from the NFS server, looks up the corresponding tag in the Virtual SCSI request list, and because this corresponding tag does not exist, the emulator drops the reply therefore emulating a SCSI abort command.

The process is shown below from box labelled “710” to “720” and finishing at “730“.

US07865663-20110104-D00007

In the patent, the above process is summed up nicely in the following paragraph.

Accordingly, a faithful emulation of SCSI aborts and resets, where the guest OS has total control over which commands are aborted and retried can be achieved by keeping a virtual SCSI request list of outstanding requests that have been sent to the NFS server. When the response to a request comes back, an attempt is made to find a matching request in the virtual SCSI request list. If successful, the matching request is removed from the list and the result of the response is returned to the virtual machine. If a matching request is not found in the virtual SCSI request list, the results are thrown away, dropped, ignored or the like.

So there we have it, that is how VMware’s patented SCSI Protocol emulation allows SCSI commands not supported natively by NFS to be honoured, therefore allowing applications dependant on Block based storage to be ran successfully within a VM where its VMDK is backed by NFS storage.

Let’s recap what we have learned so far.

1. The SCSI Commands, abort & reset have no equivalent in the NFS protocol.
2. The VMware SCSI Emulation process handles SCSI commands not supported natively by NFS thanks to the Virtual SCSI Request List.
3. Guest Operating Systems and Applications running in Virtual Machines on ESXi issue native SCSI commands to the NTFS volume, which is presented to the VM via a VMDK and housed on an NFS datastore.
4. The underlying NFS protocol is not exposed to the Guest OS, Application/s or Virtual Machine.
5. The SCSI Commands, abort & reset are emulated by the hyper visor through removing these requests from the Virtual SCSI emulation list.

In part two, I will discuss Forced Unit Access (FUA) & Write Through.

Integrity of Write I/O for VMs on NFS Datastores Series

Part 1 – Emulation of the SCSI Protocol
Part 2 – Forced Unit Access (FUA) & Write Through
Part 3 – Write Ordering
Part 4 – Torn Writes
Part 5 – Data Corruption

Nutanix Specific Articles

Part 6 – Emulation of the SCSI Protocol (Coming soon)
Part 7 – Forced Unit Access (FUA) & Write Through (Coming soon)
Part 8 – Write Ordering (Coming soon)
Part 9 – Torn I/O Protection (Coming soon)
Part 10 – Data Corruption (Coming soon)

Related Articles

1. What does Exchange running in a VMDK on NFS datastore look like to the Guest OS?
2. Support for Exchange Databases running within VMDKs on NFS datastores (TechNet)
3. Microsoft Exchange Improvements Suggestions Forum – Exchange on NFS/SMB
4. Virtualizing Exchange on vSphere with NFS backed storage?

Example Architectural Decision – vSphere Path Selection Plugin (PSP) for IBM SVC Storage

Problem Statement

What is the most suitable multipathing policy when using IBM SVC storage?

Requirements

1. Ensure maximum performance and availability for vSphere storage
2. Ensure storage performance is as consistent as possible

Assumptions

1. IBM SVC Storage which is Active/Active
2. VAAI is supported and enabled

Constraints

1. Solution must be supported

Motivation

1. Ensure optimal performance and redundancy
2. Minimize Latency

Architectural Decision

Use vSphere Native Multipathing Plugin (NMP) and configure “VMW_PSP_RR” (Round Robin) as the path selection policy.

Set the default PSP to “VMW_PSP_RR” (Round Robin) for SATP VMW_SATP_SVC so all new LUNs automatically use Round Robin

Justification

1. Round Robin helps ensure minimum average latency to the storage by using all available paths
2. Ensure performance is not degraded for some/all virtual machines due to a single HBA or connection being heavily utilized
3. Using “VMW_PSP_ FIXED” requires the paths to be manually load balanced to avoid thrashing a single path
4. Using “VMW_PSP_MRU” or “VMW_PSP_ FIXED” may lead to incosistent performance across the LUNs due to some paths being more heavily used than others
5. There is no MPP currently supplied by IBM for SVC storage
6. Round Robin is a supported configuration (Note: Although not specifically listed in the Compatability Matrix)

Alternatives

1. Use “VMW_PSP_FIXED” (Default) – Fixed Pathing
2. Use “VMW_PSP_MRU”  – Most Recently Used
3. Use vendor supplied Multipathing Plugin

Implications

1. None

vmware_logo_ads

Example Architectural Decision – Storage DRS Configuration for VMFS Datastores in a vCloud Environment

Problem Statement

In a production , self service vCloud Director environment, What is the most suitable Storage DRS configuration to improve storage utilization , performance, as well as reduce administrative effort for BAU staff?

Requirements

1. Make the most efficient use of the available storage capacity
2. Maintain consistent level of storage performance
3. Reduce the risk and overhead of capacity management
4. Reduce the risk of a unintentional or otherwise DoS event caused by self service

Assumptions

1. vSphere 5.0 or later
2. VMFS 5 Datastores which are Thick Provisioned
3. Deduplication is not in use
4. VAAI is supported by the array and enabled across the vSphere environment
5. All datastores in each respective Datastore clusters reside on the same RAID type with similar spindle types and count
6. All datastores are presented to all hosts within the cluster
7. Array level snapshots are not in use
8. IBM SVC Storage is being used
9. vCloud Director 5.1 or later
10. Storage I/O Control is enabled at set to 30ms

Constraints

1. IBM SVC storage does not currently support VASA (VMware API for Storage Awareness)

Motivation

1. Ensure production storage performance is not negatively impacted
2. Minimize the vSphere administrators workload where possible

Architectural Decision

Set the DRS automation setting to “Fully Automated”

  • Set “Utilized Space” threshold to 80%
  • Set “I/O latency” to 15ms
  • I/O Metric Inclusion – Enabled

Advanced Options

  • No recommendations until utilization difference between source and destination is: 10%
  • Evaluate I/O load every 8 Hours
  • I/O Imbalance threshold  4

Justification

1. Setting Storage DRS to “Fully Automated”  ensures that the administrator does not need to be concerned with initial placement of virtual machines as this will be dynamically and intelligently determined and executed

2. “XCOPY” is fully supported for Block based storage, as such, any Storage vMotion activity is offloaded to the array therefore removing the I/O overhead on the compute and storage fabric.

3. Where a significant I/O imbalance is detected by SDRS, the vSphere administrator is not required to take any action, Storage DRS can identify and remediate issues which fall outside parameters (which are determined by the VMware Architect) automatically. This improves the efficiency of the environment, and reduces the involvement of BAU.

4. SDRS provides valuable “initial placement” for new virtual machines which will help avoid a situation where datastores are unevenly balanced from a capacity perspective in the first place, therefore reducing the chance of virtual machines requiring migration.

5. Setting the “No recommendations until utilization difference between source and destination is” to 20% ensures that SDRS does not move virtual machines around where significant benefit is not realized  This prevents unnecessary Storage vMotion activity on the disk system, although this is offloaded from the host to the array, the I/O still may impact production performance for workloads on the same disk system.

6. Setting the “I/O Imbalance threshold (5 Aggressive / Conservative 1 ) to “4” (2nd most aggressive)  ensures that I/O imbalance should be addressed before significant imbalance is experienced by the end users. This level of “ aggressiveness” is acceptable as the Storage vMotion can be offloaded (via VAAI “XCOPY” primitive  and has almost zero impact on the host.  Setting this to “5” may result in minor I/O imbalances being corrected, at the cost of a Storage vMotion and as a result the impact of the more frequent Storage vMotion activity may negate the benefit of the I/O balancing.

7. Storage DRS will address I/O imbalance across the datastore cluster if the latency meets or exceeds the set value of 15ms (the default) and in the event of latency increasing during peak times to >=30ms , Storage I/O Control will ensure fair acess to the storage.

Alternatives

1. Use “No Automation (Manual Mode)”
2. Not use Storage DRS

Implications

1. When selecting datastores for the datastore cluster, having VASA enabled allows the “System Capability” column to be populated in the “New Datastore Cluster” wizard to ensure suitable datastores of similar performance, RAID type and features are grouped together. VASA is currently NOT supported by SVC, as such the datastore naming convension needs to accurately reflect the capabilities of the LUN/s to ensure suitable datastores are grouped together.

vmware_logo_ads