One of our customers, an important company operating in the public sector, requested HP DL380 Gen8 Server Raid data recovery following a partition corruption containing 5 virtual machines managed and created with the Debian-based Proxmox system.
Proxmox VE is the alternative Open Source solution to VmWare for managing virtualization on servers. It is based on KVM virtualization and on LXC containers capable of managing virtual machines and storage, but also virtualized networks and HA Clustering.
The DL380 system (Gen8) contained 12 disks: 2 SAS disks in raid 1 for managing the Linux Debian operating system and 10 SATA disks in raid 5 used as storage for virtual machines.
After a reboot the system started in Emergency Mode reporting problems to the Logical Volume containing the virtual disks.
Our data recovery process has developed in two distinct directions.
The usual approach involves a copy of all the discs in RAW mode to guarantee the integrity of the original media and to resolve any situations of damage in the form of bad sectors.
The images of the original disks were analyzed manually to determine the correct stripe size and order of the disks to allow a reconstruction of the raid and the extrapolation of the LV (Logical Disk) in crisis.
The operation was successful and the result obtained allowed us to recover data from virtual machines.
Another approach was based on the physical machine. The logical disk containing the virtual machines occupied a unit called “/ dev / sdb1”
With the linux command #dd / dev / sdb / dev / sdc conv = noerror, sync we have cloned the data on our external disk (dev / sdc). In this case it was the raid card of the system that dealt with the consistency of the raid. Also with this method the same result was obtained, a total and successful data recovery
Reference for- Data Recovery Cost