Questions, answered.

maxRAID is designed for data‑center‑class TLC SSDs rated at 1 drive write per day (DWPD). These drives are our baseline for calculating wear and capacity.

Any SSD interface — SATA, SAS, NVMe — and fabric targets such as iSCSI and NVMe‑oF. Most customers choose local drives for the best value and performance.

It depends on the kind. Some high‑endurance SSDs simply add over‑provisioning, which reduces maxRAID’s benefit. Others use higher‑grade flash, and deliver the full benefit of maxRAID.

Yes. Data‑center QLC SSDs perform well with maxRAID. They have QLC limitations, but maxRAID mitigates most of the performance and wear issues — with far less wear, QLC may handle workloads that typically require TLC.

maxRAID suits them well. Some newer, larger SSDs map blocks larger than 4K. Because maxRAID writes large aligned blocks, the larger indirection unit doesn’t affect performance or wear.

Yes. Consumer SSDs often have higher read latency, lower endurance and inconsistent performance due to pSLC caching. maxRAID mitigates many of these issues — in some cases consumer SSDs can even outperform data‑center SSDs in a parity array.

Single SSD, RAID‑0, RAID‑1, RAID‑5 and RAID‑6 — plus 3‑ and 4‑drive parity for even greater protection.

RAID‑5 is the most common choice: space‑efficient, fast, and it writes less in total than RAID‑10. RAID‑6 is nearly as fast — modern x86_64 processors hide most of the parity overhead — and survives more than one concurrent drive failure.

With maxRAID they offer no additional protection, capacity or performance over RAID‑5 or RAID‑6, so we don’t recommend them.

maxRAID is a kernel module, so we support specific kernels: every x86_64 production kernel from Red Hat for RHEL 8 and 9, select Rocky Linux and AlmaLinux releases, and all server “‑generic” x86_64 kernels for Ubuntu LTS 22.04 and 24.04. Other distributions are added based on demand.

Yes. We work with OEMs and major cloud providers to include their custom kernels in our build set.

We recommend about 80%. Performance stays predictable as the array fills, and keeping some free space gives maxRAID room to work — for the best speed and the longest drive life.

Anywhere you use a block device. It works with file systems, LVM and partitions (most customers use LVM).

Yes. maxRAID is swap‑safe and supports swap files, partitions and logical volumes.

Not natively. Use LVM or a snapshot‑capable file system on top of maxRAID.

Yes. Writing blocks of all zeros or all ones is also treated as a trim, so space is reclaimed even without discard support in the client or hypervisor.

The device can present either 512‑byte or 4K blocks, and can be configured to issue only 4K I/O to SSDs that require it.

Yes. Export a maxRAID device — or portions of it via LVM — over iSCSI or NVMe‑oF. maxRAIDOne adds NFS, SMB and S3 with a web interface and API.

The trial lasts 120 days from installation and supports arrays up to 100 TB. If you need more, contact us.

Run maxRAID Tools. The non‑destructive device scan measures how compressible your data is and how full your devices are.