High Availability Storage Systems

Prepare the environment by provisioning three virtual machines in VirtualBox with Rocky Linux, configuring bridge and internal networks, assigning static IPs, enabling passwordless SSH, and creating snapshots.

Explore how high availability minimizes downtime through redundancy, failover, replication, load balancing, and monitoring. Understand SLA guarantees, MTTR, MTBF, MTTF, and the bathtub curve to optimize storage resilience.

Explore how the logical volume manager (LVM) overcomes traditional partitions by pooling physical volumes into volume groups and carving flexible, resizable logical volumes with snapshots, extents, and clustered setups.

Practice creating lvm on two disks with physical volumes, forming a volume group and a logical volume, building an ext3 filesystem, mounting, and verifying with pv and lv displays.

Extend storage with hands-on LVM: create a new physical volume, extend vg00, grow data LVM to 3 gb, and resize the filesystem online.

Shrink an lvm-backed file system offline by unmounting, checking free space with df -h, running e2fsck -f, and resizing to 2 gb, then remount; x3 file system resize is offline.

Explore backing up and restoring LVM metadata with vgcfgbackup and vgcfgrestore, view backups with vgcfgrestore --list, and recover VG00 by reloading metadata, scanning, activating, and reformatting logical volumes.

Learn how raid, a data storage virtualization technique, combines multiple disks into a single unit to boost performance, capacity, and fault tolerance, with software, hardware, onboard options, and common levels.

Configure a raid 1 array with mdadm, partition disks, create an xfs filesystem, mount to /mnt/raid_data, and enable monitoring and mail alerts.

Discover how raid variants respond to failure—from raid 0's no redundancy to raid 1 mirroring and raid 4/5 trade-offs—plus hot plugging enables uptime during maintenance.

Simulate a raid 1 failure, mark a disk faulty, remove it, then repair by adding a spare disk to mdadm and monitor recovery and data sync.

Add a hot spare disk to a raid array using mdadm, enabling automatic failover and reduced downtime. Monitor status and verify email alerts.

Explore iSCSI architecture, including initiator, target, and LUN roles, IQN naming, ISNS discovery, LIO targets, and file IO versus block IO with sparse and non sparse images.

Restore node one to the first snapshot and set node three as the target storage system, with all nodes in the same state and no RAID or LVM configuration yet.

Configure an iSCSI target using the targets cli, create backstores (block and file io), map LUNs, set ACLs and CHAP authentication, and verify the target starts listening on port 3260.

Install and configure the iSCSI initiator on multiple nodes, set initiator names and CHAP authentication, then discover and log into targets to use the connected block devices.

Explore DRBD, a Linux distributed replicated block device for real-time network mirroring. Learn primary and secondary roles, synchronous protocols, failover, and direct point-to-point connections.

Prepare a two-node environment by installing drbd, disabling se linux, configuring firewall ports for drbd, and creating 1 gb lvm volumes on both nodes for high-availability storage.

Simulate a split-brain scenario in a high-availability storage system using DRBD, promote a node to primary, copy data, and decide which side to keep and which data to discard.

Simulate a split brain in a high-availability storage system using drbd, observing primary, secondary, and unknown statuses. Decide which node becomes primary and which data to discard.

Configure DRBD 9 dual primary mode with a shared disk cluster file system (GFS2 or OCFS2) and LVM storage to achieve high availability across two nodes.

Master multipath io to achieve redundancy and load balancing across fibre channel, iSCSI, or SAS storage, with device mapper and active-passive or active-active modes.

Set up multipath for an iSCSI target by installing device mapper multipath, configuring multipath.conf with user-friendly names, and starting the service to verify paths and aliases.

Configure a two-node drbd setup for synchronous block replication. Create resource files, load drbd, initialize metadata, and promote node one as primary during synchronization.