Applies to: BLAZE
Summary
All BLAZE appliances have two internal storage configuration options for units with multiple drives: No RAID or RAID 1. Some models also offer RAID 5 and 6 as redundancy options. In a normal state, both drives are being written to at the same rate and are mirrored.
This article explains how the appliance behaves during a drive failure and RAID rebuild, and any temporary performance impact during this process.
Sequence of Events
NOTE: This timeline is based on a BRB-X1610SE appliance with two 4 TB HDDs
configured as a single RAID 1 volume.
Drive Failure
When a single disk fails, recording continues uninterrupted. A Storage Error alarm notification appears in the notification panel. If rules have been configured, they trigger actions, including email notifications and HTTP commands to other systems.
BLAZE Health Monitoring reports the disk is unavailable.
Navigate to Settings > Servers > Storage to view reports that the RAID volume is degraded.
The event is recorded in the Event log.
Drive Replacement
For BRB appliances:
- Power down the BRB appliance and disconnect power.
- Remove the cover of the appliance and replace the faulty drive.
For BRR-X and BRR-P appliances:
- Swap the drive. No need to power down the appliance; when configured for RAID, their drives are hot-swappable.
- If the appliance is not configured for RAID, power down the BRR appliance and replace the faulty drive.
RAID Rebuild
BLAZE immediately initiates a RAID rebuild. It is a full rebuild, meaning it destroys existing data on the new inserted drive. Recording continues uninterrupted during the rebuild, but does slow down the process. Shown in Health Monitoring, the disk Read and Write data rates reflect the intense transfer of data from disk #2 to disk #1 to rebuild the RAID.
For example, assuming a BRB with 8 cameras, recording at around 20-50 Mbps, with two 4TB drives configured for RAID 1 and full of recorded video, the rebuild time was around 12 hours. The estimate decreased to around 7 hours when recording was temporarily disabled.
During a RAID rebuild you are vulnerable to a second drive failure, which would cause total loss of data (assuming RAID 1 or 5). You can accelerate the rebuild process by reducing the load on the appliance.
Strategies for achieving this include:
- Pause recording.
- Change the schedule.
- Use Motion / objects / events recording instead of Record always.
- Reduce the framerate or resolution.
- Minimizing the number of searches, replays, and exports.
Normal State
Upon successful completion of the RAID rebuild, the Rebuild warning disappears.
The Read/Write statistics return to normal. In the example below, Writes peak at around 10 MB/s compared to over 100 MB/s witnessed during the rebuild.
The Event log records the moment of the disk failure (11:57 AM in the example below), and the automatic system-initiated RAID rebuild is captured in the Audit log (2:43 PM in the example below).
Levels of Redundancy
| LEVEL | DESCRIPTION |
| NO RAID | No redundancy. The failure of any disk causes all data on all disks to be lost. Full disk space utilization even with different sized disks. This is the default storage configuration for BLAZE appliances. |
| RAID 1 | Mirroring. Data is duplicated across both disks, so that it can survive the failure of one disk. RAID 1 results in 50% disk utilization. Requires an even number of disks. |
| RAID 5 | Striping with single parity. Survives a single disk failure. Only the space of one disk per volume is lost to provide redundancy, so for an 8-disk volume this offers 87.5% disk space utilization. Requires 3 or more disks per volume. |
| RAID 6 | Striping with dual parity. Survives two concurrent disk failures. The space of two disks per volume is lost to provide redundancy, so for an 8-disk volume this offers 75% disk space utilization. Requires 4 or more disks per volume. It is more resilient that RAID 5, but rebuild time is longer due to the added computational complexity. |