Linstor, ready for Kubernetes

The recent session with LINBIT during The IT Press Tour was good as we had time to dig into LINBIT’s Kubernetes strategy essentially covered by Linstor, a dedicated storage solution for Kubernetes. As the DNA of the company, the product is open source under the GPL license and goes beyond Kubernetes also supporting OpenNebula, OpenStack and OpenShift but it is above all a generic storage volume management for Linux. It’s obvious that the product targets persistent volume for stateful workloads.

The architecture resides on multiple services with active/passive manager named the Linstor controller, one active at a time, and a series of agent, called Linstor Satellite, deployed on each node in the cluster under the controller management. The Linstor Controller is managed via API or CLI. Linstor also integrates DRBD for data resiliency at the block level.

For Kubernetes, Linstor exposes CSI and LINBIT also develops an operator named Piraeus with Doacloud. This project is a sandbox CNCF project. As an analogy it plays the same role as Rook for Ceph. Definitely one of the few right storage management service to associate for each Kubernetes cluster deployment.

Read More

Atempo has made lots of progress

Atempo joined the recent IT Press Tour in Lisbon, Portugal, for the 5th time. As the company did a lot of things during the last few months, it was the right opportunity to take time with 2 key executives instrumental in the success of Atempo: Ferhat Kaddour, VP Sales and Alliances, and Louis-Frederic Laszlo, VP Product Management.

Founded 30 years ago in 1992 in France under the name Quadratec, Atempo is a reference in data protection and data management software, with more than 2000 customers and more than 200 employees. The company sells its products exclusively via a strong channel built for a few decades to address some vertical use cases or operate as a generic answer.

Atempo has had several lives as an independent entity, under the ASG - Allen Systems Group - umbrella and back as an independent firm since 2017 when Luc d’Urso acquired Atempo from ASG via his kickstart legal entity. On the history timeline, in 2006, Atempo has acquired Storactive to become Lina, in 2017, Atempo picked Synerway, another french data protection ISV, and finally in 2019 swallowed Lima and has initiated an AI lab named Nextino. All of these consolidate Atempo as a strong data protection and management company.

As of today the product line is split between data protection and data management from SMB to large enterprises aligned with vertical industries and domains. The product family is represented by the historical strong software Tina aka Time Navigator, Lina or Live Navigator, the NEW Continuity and Miria.

Let’s dig a bit into Miria. As we know the solution is super open, probably one of the most open in its category, supporting a wide variety of file servers from NAS to parallel file systems, access methods, private or public clouds via S3 and related object stores and finally tape.

One of the recent few announcements was the partnership with Quantum to couple Tina with Quantum DXi and the Panasas one around Miria. WE saw also agreements with Storj, CIQ around Rocky Linux, DataCore or Seagate. Some of us also watched the OVHcloud cold archive service built on IBM tape library and 3592 tapes controlled and managed by Atempo Miria.

Miria was also listed in the recent Coldago Map 2022 for Unstructured Data Management (LINK) as a challenger without any leader in the domain.

The team has shared ambitious elements for 2023 and beyond around data immutability, data integrity, cost control and sustainability. Among these we saw fast-scan and block deduplication for S3 and some container flavors.

Read More

LINBIT, the reference in Linux Data Protection

LINBIT, the reference in Linux data protection, joined the recent 47th IT Press Tour organized in Lisbon, Portugal. It was the opportunity to get an update about DRBD, the famous Distributed Replicated Block Device engine, directly from the CEO and founder of LINBIT, Philipp Reisner, who started this project during his studies more than 20 years ago.

Since that, the company has launched another product named Linstor, with a portion of the company name in the product name. Perhaps a recognition of the disconnect between a recognized product and an unknown company, people having difficulties to associate the two. will publish an article later about Linstor but also LINBIT SDS, HA and DR.

LINBIT was founded in 2001 in Austria without any VC and employs around 35 people today. The company is profitable and fully owned by its founder.

I mentioned DRBD as a famous solution by the simple fact that the product is deployed more than 2 million times. It is used by large well known brands like Amazon, Apple, Cisco, Google, IBM, Intel, Oracle, Red Hat or Yahoo!.

So let’s cover today DRBD as the preferred block level replication for Linux. As a block oriented engine, the copy engine works above any block device, physical or logical, and below the file system layer. A block device could be a single partition, a LVM volume, a MD volume with 0, 1, 4, 5, 6 or 10 assembly mode, a VDO (Virtual Data Optimizer from Permabit acquisition) entity or a zVol from ZFS with different flavor. DRBD supports any initiators and targets supported by Linux such LIO, STGT, SCST, Ietd, iSCSI and of course NVMe-oF and especially TCP. It offers synchronous and asynchronous mode and up to 32 replicas with version 9.x on any TCP/IP network.

The mode can be mixed between one source and its targets. Multiple primaries are also possible for cluster needs like GFS or OCFS2. Cascading is not supported yet but as multiple replicas are possible users can build interesting configurations but still add load on the primary. Users can even configure the engine to be diskless at the source copying memory data to remote devices. The engine implements partial resync to only consider blocks not already present or synchronized. The replication can work in multiple directions with only one master source at a time except on some specific use cases such cluster mentioned above or live migration of VMs. During a failover, the unmount of file system then detach of volume triggers the source to release its primary role and vice-versa, as soon as a volume is locally attached and the file system is mounted, the initiator is set and the replication direction is decided. But LINBIT protects this and only one source can be active at a time preventing data corruption, except the 2 cases listed above.

The team has launched a Windows flavor, currently in 1.0 released in Q1/2022 supporting Windows 7SP1, Windows 10, Windows Server 2008, 2016 and 2019.

DRBD 9.0 and 9.1 can be downloadable here, the 9.2 announced October 10 is also available there.

Again a very good deep session presented by the master Reisner.

Read More

Cleondris, the perfect companion for NetApp data protection deployments

Cleondris, a Swiss software company developing add-ons for NetApp systems, joined for the first time The IT Press Tour, the 47th edition, a few days ago in Lisbon, Portugal. It was the perfect opportunity to meet Christian Plattner, CEO and founder, for me it was the second as I discovered the company and met Christian at NetApp Insight 2018 in Las Vegas.

The company, founded in 2006 in Zurich, Switzerland, was launched with the mission to offer special software for ONTAP and to deliver these solutions for each NetApp product iteration like Clustered ONTAP, SnapDiff or SnapMirror. Today, 3 products are available: SnapGuard, Data Manager and Cleondris HCC.

• SnapGuard targets ransomware with the capability to potentially block usual requests via FPolicy module on ONTAP. It implements a sort of firewall for NFS and SMB file access.
• Cleondris Data Manager (CDM) is an orchestrator for NetApp snapshot mechanism and controls how you backup and restore any ONTAP-based filer.
• HCC is a special add-on for SolidFire when users deploy a cluster. It facilitates failover of VM-based workloads hosted in such clusters. This product is on NetApp price list but it is sold only to users with already deployed SolidFire, outside of these, the NetApp product is end of life.

All Cleondris solutions rely on the Cleondris Appliance, a VM-based appliance, running CentOS, PostgreSQL and other Cleondris specific Java Software.

The appliance supports VMware OVA, Linux RHEL/CentOS and physical server and is potentially available on demand on AWS and as a Docker container. AWS will be officially available soon and the product already supports Amazon FSx for NetApp.

CDM offers a comprehensive GUI to schedule SnapMirror and is capable of restoring VMware objects with a fine granularity: Full VM restore, VM clone, VMDK restore, VMDK attach to offer file-level restore, Datastore clone and DR with HCC. Of course NDMP control is integrated as well. Cleondris is able to monitor all data movements coupled with vCenter and both source and target storage devices. As Cleondris Data Manager sits in the back and monitors all Snapshots and SnapMirrors jobs, CDM is the right tool to globally see all protection tasks and trigger any restore. CDM is capable to index all files on any ONTAP volumes and thus know perfectly if all files are protected by a Snapshot or SnapMirror jobs and present in a “backup” image, tracking also versions. Cleondris belongs to the small list of official partners for SnapDiff v3. In terms of scalability, 1 Cleondris instance is able to support 10 of thousands of VMs and hundreds of NetApp clusters.

SnapGuard illustrates perfectly the pioneer role of Cleondris in ransomware protection for NetApp systems. It is available for FAS, AFF, Select, CVO and Amazon FSx for ONTAP. This module is able to scan 300k FPolicy messages and 16k ransomware checks per second and CPU core. The effect is powerful with immediate blocking mode with export to SIEM solutions for advanced analysis.

Cleondris is licensed per ONTAP or ESXi node whatever is the volume of data or number of users or number of ONTAP volumes.

Clearly a must have for any ONTAP user as the Cleondris suite of tools enhanced NetApp environment, already demonstrated at hundreds of sites. It was a very good dense session presented by Christian Plattner.

Read More

47th edition of The IT Press Tour organized in Europe

The IT Press Tour is back in Europe for the 47th edition globally and we'll be in Lisbon, Portugal, for a few days next week. This tour will be dedicated to IT Infrastructure, Data Management and Storage and we'll 6 hot European companies:

• Atempo, France, the European leader in data protection,
• Cleondris, Switzerland, developer of key solutions on top of NetApp,
• Linbit, Austria, the reference in Linux data resiliency,
• NGX Storage, Turkey, a growing unified storage player,
• Nuclia, Spain, a emerging actor in content indexing levering AI,
• and Tiger Technology, Bulgaria, a reference in hybrid cloud file storage.

I invite you to follow us on Twitter with #ITPT and @ITPressTour, my twitter handle and @CDP_FST and journalists' respective handle.

Read More