Are standard reservationless ZFS Snapshots used?

Yes, SoftNAS uses reservationless ZFS snapshots.

Can I login to SoftNAS UI using another system user?

Yes, you can create any other user on softnas and still be able to login to the UI with full privileges. However, you must first add the user to the 'softnas group'.

Does SoftNAS provide High-Availability (HA)?

Yes. SoftNAS patented SNAP HA® delivers a low-complexity solution for high-availability clustering that is easy to deploy and manage. A robust set of HA capabilities provides constant access to your data even when datacenter infrastructure failures such as datacenter, server, network and storage system failures to keep your business running without storage downtime.

Does SoftNAS work as a Virtual Storage Appliance (VSA)?

Yes, SoftNAS Cloud NAS is delivered as a Virtual Storage Appliance and runs as a virtual machine on Amazon EC2, Microsoft Azure or VMware vSphere.

Does SoftNAS support SCSI-3 persistent reservations for clustered hosts?

Yes, SoftNAS supports SCSI-3 persistent reservations.

Does SoftNAS use the standard ZFS ZIL and L2ARC mechanisms for write/read acceleration (respectively)?

SoftNAS leverages the local ephemeral SSD storage and RAM in the instances for the caching system. It’s a great part of our solution because we can enable any combination of compute, caching and attached storage based on the wide variety of the cloud offerings. We encourage customers to customize based on their needs. As new instances come out or improve, we immediately encourage customers to make the change! This is impossible to achieve with the conventional on-premise storage model.

Do you need to be a storage administrator to manage SoftNAS?

Not at all! SoftNAS StorageCenter™ is the easy-to-use, intuitive GUI to manage SoftNAS Cloud NAS. The management software is accessed via your web browser and provides simplicity for creating standard mounts, including NFS, CIFS, and iSCSI storage volumes, which can be pooled into RAID backed arrays to customize redundancy and performance as well as setting advanced features.

During data copy on replication, if 90% of the copy is done and the link breaks, will it start from zero?

If the copy replication is broken the, upon restart the copy resumes from the last known good snapshot. DeltaSync takes place after a fail over event which only copies the differences since the node was active.

How Copy on Write (COW) file system and Data Integrity work?

ZFS snapshots are based on tree snapshots which are stored considerably different than how block snapshots were traditionally done. The Copy On Write file system creates a map as to where the changed blocks are stored instead of traditional method of what was overwritten. This gives the ability to have unlimited snapshotting and facilitates the benefit of the little or no space that current ZFS snapshots consume. The below Article from Jeff Bonwick of Sun gives a great visual depiction of how tree snapshots are structured.

https://www.snia.org/sites/default/orig/sdc_archives/2008_presentations/monday/JeffBonwick-BillMoore_ZFS.pdf

How does the HA piece work, exactly?

SNAP HA™ monitors all critical storage components, ensuring they remain operational. When there is an unrecoverable failure in a system component, another storage controller detects the problem and automatically takes over ensuring no downtime or business impacts occur.

You can refer to the below section for more information:

How is encryption key management handled?

We leverage AES 256 via LUKS. You can find more information on this subject here: https://www.softnas.com/docs/softnas/v3/html/security.html

How is ZIL utilized by SoftNAS?

The ZFS Intent Log (ZIL) absorbs bursts or steady streams of incoming, synchronous write operations by rapidly recording the writes in high-speed, mirrored SSD storage, then returning quickly to the caller. This makes synchronous writes operate at maximum speed. Later, when there is less I/O overhead, the writes which are cached in the ZIL are written to long-term pool storage. The ZIL is flushed every 5 seconds

On AWS some customers use a small (5 GB) Provisioned IOPS SSD (never direct attached ephemeral SSD) to absorb massive write bursts. It does not need to be large because it only buffers a small timeframe of data.

How to create an iSCSI volume for Windows Failover Cluster?

Windows Failover Cluster requires a special block size and must be set manually:

Use iSCSI as the storage protocol on SoftNAS, iSCSI Initiator on the MS servers.
You have to create the pools via the GUI, but create the volumes via cli to get this.

Example below using a 40G volume and thick provision:

zfs create -V 40G -b 4k pool01/iscsi01 (crate the vol with a 4k block size)

zfs get volblocksize pool01/iscsi01 (verify)

zfs set reservation=40G pool01/iscsi01 (set the reservation size)

zfs get reservation pool01/iscsi01 (verify)

zfs list (verify)

3. Format the volumes to have 4k block size for NTFS. (ZFS uses 128k by default)

How to find the directories that are occupying the most space on the system?

The below command will list the 15 directories with the highest disk usage. The -x option will make it only search the root file system and ignore any other mounted systems specially ZFS pools, as scanning them may take long time and affect the performance.

du -xh / 2>/dev/null | sort -h -r | head -n 15

If you would like to check for a certain pool, you can replace "/" with "/Your-Pool"

In a hybrid approach, is it feasible to have a 2-node SoftNAS deployment on-premises that tiers data out to (and back from) public cloud? Or do we need to replicate SoftNAS on-premises to SoftNAS in Azure/AWS?

We can put a pair of HA enabled nodes on premise for local storage and connect to Azure/AWS for cloud capacity. We do indeed enable replications and/or snap shots to the Azure/AWS hosted instances. Separate shares on Azure/AWS can be used for expanded capacity.

Is my data encrypted, and who controls the encryption keys?

SoftNAS encrypts your data while it is at-rest and also when it is in transit—we call it 360-degree encryption. SoftNAS believes that each business or organization should control the keys to their data—so only you have the encryption keys, not SoftNAS nor your underlying storage cloud platform provider.

Is standard Linux clustering used, or have you written your own stack?

Yes, we a CentOS based for our ZFS implementation (not Illumos… thank goodness).

Is there a 2-node max per deployment? Or can we scale out horizontally too?

We use two nodes for a high availability (HA) solution. The performance scales by changing the instance's type rather than node scale out. We can attach to a combination of back end storage options (block and object) for up to 16PB of capacity as well.

Is there a AES cipher that traffic passes through to encrypt data that is sent to disk via replication?

The SnapReplicate® feature provides block-level replication between two SoftNAS instances. SnapReplicate between source and target SoftNAS instances sends all data through encrypted SSH tunnels and authenticates using RSA PKI. Data is encrypted in transit using industry-standard ciphers. The default cipher for encryption is Blowfish-CBC, selected for its balance of speed and security, but you can use any cipher supported by SSH, including AES256-CBC.

Is there any bench marking tool you are using and if there is any specific configurations?

Feel free to use the bench marking tool that you want that allows you to best represent your exact work load Read/Write percentage.

What are the RPO and RTO times for SoftNAS Cloud NAS?

RPO is an acronym for Recovery Point Objective which refers to the most recent point in time for which you will be able to recover. SoftNAS has a default RPO of 60 seconds. i.e. After the initial synchronization of data between two SoftNAS nodes, SnapReplicate will sync changes between SoftNAS appliances every 60 seconds. Provided the amount of changed data over the past 60 seconds completes its transfer prior to the next sync your RPO is equal to that target. (The network and compute may be adjusted based on actual needs to keep RPO as low as possible).

RTO is an acronym for Recovery Time Objective or the amount of time required to recover from a failure. SoftNAS employs various mechanisms to enable customers to quickly recover from accessibility issues. This recovery ranges from manual efforts to automated HA events. Depending upon the architecture of the system, the nature of the ‘event’ RTO can be as low as 30 seconds.

What happends if my license expires on my SoftNAS deployment?

There is a 7 day grace period where everything will continue to function to allow you to purchase a new license without interruption.

Once the 7 day grace period expires, you should expect the following:

Scheduled snapshots will stop taking place.
Replication will stop.
HA will deactivate but the VIP will remain associated to the source node so that clients can continue to access the volume mount points.
You will lose Storage Manager administrative access.
Only two pages will function in Storage Manager.
1. Licensing console. This allows you to enter your new license information once you have it procured.
2. Software Updates console.

What happens to my data if the SoftNAS instance is stopped?

Your data remains intact and undisturbed. When SoftNAS is not running, NAS clients connecting via NFS, iSCSI, or CIFS will not have access to NAS services, which will be restored when the SoftNAS is powered back on. If data is in production with SNAP HA®, a failover would occur and the data would still remain available.

What versions of SMB and NFS are supported?

SoftNAS supports SMB v3 and NFS v3 and v4.

Where is the GUI run from?

Our admin interface, Storage Center, is web based. It’s on either node and it unifies multiple SoftNAS deployments.

Why is Open Source a good foundation for SoftNAS?

Open Source software can be enhanced by a wealth of programmers, which means bugs are easier to find and fix, and the software is much more secure. Although SoftNAS is not Open Source, it is based on trusted Open Source industry standard software–Linux and ZFS–which allows an open and free architecture, so your data is never locked into one platform.