RAID-5 with 3 failed disks on a bank BRSA archive server · 99.4% recovery in 48 hours
On the archive server of a mid-to-large-scale Turkish bank, a 4× 4 TB Enterprise SAS RAID-5 array · first one disk died and a rebuild started, then 2 more disks died. 7 TB of BRSA compliance records were at risk. A 99.4% recovery in 48 hours with a clean room + mathematical reconstruction.
01 The Challenge
Monday 09:14 · the bank IT team called DSET: the back-office archive server is not booting · the LSI MegaRAID 9361 controller reports 'array failed.' A 4× HGST Ultrastar SAS RAID-5 · 3 disks not lighting up green. The last backup was 60 hours earlier · a BRSA audit in 5 days · proof of data integrity essential. Within the 7 TB are customer transaction records, the MASAK reporting archive, internal audit reports. An on-site rebuild attempt could have burned the 4th disk too.
02 DSET's Approach
T+0 · Triage by phone
The DSET data recovery team assessed the situation within 9 minutes. The bank was given the directive: 'no rebuild attempts · do not remove the disks from their slots · ventilated transit ESD packaging.'
T+4h · Clean room transfer
The disks were taken by a special courier to an ISO 14644 Class 100 clean room. PCB damage + ESD pre-checks were done.
T+8h · Disk-by-disk diagnosis
Disk-1: firmware read instability · disk-2: head crash · disk-3: spindle motor problem · disk-4: healthy. A raw image of each was taken with PC-3000 + DeepSpar.