⚠ Do now
Free tool: Array Passport
Fill it in while the array is healthy: the RAID level, which drive sits in which bay, every serial number, the software version and who to call. It prints one A4 page for the cabinet.
Last checked
If this array fails
Store this inside the cabinet or with your offline records. It contains no passwords or keys.
wildfiredata.com.au · 1300 806 557
More about this tool
The passport is a form that runs entirely in your browser. You type what you know, it builds the printed sheet as you go, and nothing leaves your computer unless you save or copy it yourself. It suits Synology, QNAP, TrueNAS, Unraid, Windows Server and hardware RAID controllers.
Seven short steps: device, array, drives bay by bay, protection, access and encryption, people, and what to do if it fails. Skip anything you do not know. Blank lines still print, so you can fill them in by hand. With the NAS settings open, it takes about ten minutes.
On a Synology, a QNAP running QTS or another NAS built on Linux software RAID, your IT provider can paste the output of mdadm --detail. The form reads the RAID level, the drive count and the member list from it, and ignores anything it cannot read. mdadm lists partitions, not bays, so the drive lines still come from you.
Print one A4 page, black on white. Download a copy as a small .json file and open it again next year with Open a saved copy. Copy as text gives a plain version for an email or a ticket.
One page holds up to eight drives. Bigger arrays, up to 24 drives, and very long answers print in smaller type, and the form tells you if the sheet would run onto a second page.
A RAID array spreads your files across several drives in a fixed pattern. To read them back, the software has to know the RAID level, which drive held which position, the chunk size (how much data goes on one drive before the next one takes over) and how the parity is laid out. While the array is healthy, the NAS keeps track of all of that for you.
On a NAS that uses Linux software RAID, each member drive carries a small record called a superblock. It holds the RAID level, the layout, the chunk size, that drive's role in the array and an event counter that goes up every time the record is updated. When mdadm puts an array back together, it uses that counter to spot members that dropped out or fell out of step. Hardware RAID controllers keep a copy of the layout as well. When a Dell PERC controller finds drives whose configuration it does not know, or that does not match its own, it reports a foreign configuration.
Those records describe the array as it is now, and they live on the same drives that are failing. A pool that was re-created or initialised, a rebuild that stopped halfway or a replaced controller can leave them missing, overwritten or out of step with each other. mdadm can even be told to assemble an array when the records on some members look out of date, and its manual warns that an array which needs that may contain data corruption.
That is when a recovery lab has to work the settings out from the raw data: the drive order, the chunk size, the parity pattern and which members were current when the array stopped. Every unknown adds testing, and a wrong guess written back to the drives makes the job harder. A record taken while the array was healthy gives the engineer a starting point from you instead of a guess. It does not replace the lab's own checks, because drives get moved and settings change, but it gives those checks something to start from.
The metadata version matters too. Linux RAID keeps the superblock near the end of each member for versions 0.90 and 1.0, at the start for 1.1, and 4 KB from the start for 1.2. If you paste mdadm --detail into the passport, the metadata version, the chunk size and the member list go onto the sheet.
Good NAS documentation answers the questions a recovery lab asks first. It never holds the secrets that would let someone else in. These are the seven steps of the passport, and the things it will never ask for.
The form asks where the key is kept, never for the key. In the key and notes boxes it warns you if something looks like a recovery key or password, and asks you to take it out.
Without its key, an encrypted volume stays locked, for you and for any recovery lab. Knowing where the key lives, and who can reach it, matters as much as the drive order.
It writes changes to the SSDs first and copies them to the hard drives later, and Synology warns that removing the SSDs before that data is written back can lose data. Record the cache SSDs and their slots. If the NAS fails, the SSDs go to the lab with the drives.
RAID keeps running when a drive fails, but it copies deletions and ransomware straight across, so it is not a backup. ASD's Essential Eight expects organisations to test restoring data, applications and settings from backups as part of disaster recovery exercises. The date on the sheet shows when that last happened.
Keep at least two copies, and never only on the NAS it describes: a copy stored on the array is lost with the array. The sheet holds names, phone numbers and the unit's network name, so treat it as an internal document.
Keep the printout with the NAS, in a sleeve inside the comms cupboard door or the rack. Whoever is standing at the unit on a bad day has the bay order in front of them, and the sheet ends with three steps to follow if the array fails.
Keep the downloaded .json file and a second printout somewhere that does not depend on the NAS, such as your business continuity folder or a USB stick kept away from the office.
Send them the .json file, or the Copy as text version, for their client records. Then they can answer a lab's questions even when you cannot be reached.
Keep on this device saves the form in that browser only, and only while it is ticked. It is handy on your own computer. Leave it off on a shared one, and do not rely on it as your only copy.
A passport is only as good as its last update. The form sets a review date twelve months after the date captured. Any of these changes should bring it forward.
To update, use Open a saved copy with last year's .json file, change what is different, then print a fresh sheet and swap it for the old one in the cabinet.
On 29 October 2026 Synology said attackers were targeting NAS units on outdated DSM through a vulnerability it patched in December 2024 (Synology-SA-24:20). Its advisory Synology-SA-26:19 of 10 November 2026 sets the minimum safe builds: DSM 7.4.1-90080-1, 7.3.2-86009-5, 7.2.2-72806-10 and 7.2.1-69057-13, and lists no fixed build for DSM 7.2, 7.1 or 6.2.
A full build number on the sheet can be checked against that list in seconds. Our NAS ransomware check does the comparison for you, and the C3RB3R guide covers the rest.
If you look after NAS units for clients, the passport is a handover document. Fill one in when you take on a client, when you commission a new unit, and when a site passes to someone else.
mdadm --detail over SSH to capture the level, member list, chunk size and metadata version. Then check the drive lines by hand, because mdadm lists partitions, not bays.A made-up but typical example: a small Brisbane office with a four-bay Synology, filled in by its IT provider on 6 November 2026. The model, version and serial numbers are examples only.
Captured 6 Nov 2026
Review by 6 Nov 2027
| Bay | Make and model | Serial number | Capacity | Installed |
|---|---|---|---|---|
| 1 | WD Red Plus WD40EFPX | WD-EXAMPLE01 | 4 TB | Mar 2022 |
| 2 | WD Red Plus WD40EFPX | WD-EXAMPLE02 | 4 TB | Mar 2022 |
| 3 | Seagate IronWolf 8 TB | ZA-EXAMPLE03 | 8 TB | Feb 2026 |
| 4 | Seagate IronWolf 8 TB | ZA-EXAMPLE04 | 8 TB | Jun 2025 |
mdadm members: two data sets, a RAID 5 set across all four drives and a RAID 1 set across the two 8 TB drives.
A shortened version of the printed sheet. The real one adds protection, access, people and the three steps to follow if the array fails.
SHR is Synology's own RAID system. Synology says it is based on Linux RAID and divides each drive's space into smaller chunks, so drives of different sizes are not wasted. With mixed sizes, that usually means more than one RAID set on the same drives, as in this example. The IT provider chose SHR in step 02, then pasted the mdadm --detail output for both sets. The form kept SHR as the RAID level instead of replacing it with RAID 5, filled in four drives, and put both sets in the member list.
Questions
Short answers about RAID and NAS documentation, and what this form does with what you type.
Enough for a stranger to identify the unit and rebuild its layout: the make, model and exact software version; the RAID level, file system, pools and any SSD cache or hot spare; every drive by bay with its serial number; the backups and the last test restore; where any encryption key is kept and who holds it; and who to call. The Array Passport is a free template that covers all of it on one page.
It is free, with no account. It runs in your browser and makes no network calls, so nothing you type is sent anywhere. A copy only exists if you print it, download it, copy it, or tick Keep on this device, which saves it in that browser alone.
No. Record where the recovery key is kept and who holds it, never the key itself, and keep passwords in a password manager. A sheet that held the key would hand it to anyone who found the sheet. The form warns you if something in the key or notes boxes looks like a key or password.
After any drive swap, expansion, change of RAID type, software or firmware update, or move, and at least once a year. The form sets a review date twelve months after the date captured. Open last year's .json file, change what is different and print a fresh sheet.
No, it never connects to the NAS. Most answers are on the label underneath and in the NAS settings: Storage Manager on a Synology, Storage & Snapshots on a QNAP, or Storage on TrueNAS. On a Synology, a QNAP running QTS or another NAS built on Linux software RAID, your IT provider can paste mdadm --detail output to fill in the RAID level, drive count and member list. For a hardware RAID controller, write the controller model and firmware in the version box.
The numbers in the text link here. Last checked ; source 8 updated 11 November 2026.
Already in trouble?
Do not rebuild, repair, initialise or swap drives. Label each drive with its bay number, keep any cache SSDs with them, and talk to our Brisbane lab first.