Inventory and status of zare systems at CoF
Status as of 14 September 2026. All figures measured directly from both servers that day. Unlisted; not linked from anywhere.
The short version
CoF's video archive lives on two servers, zare4 and zare5: six storage volumes, about 166 TB of usable space.
- All six volumes are online with full two-drive redundancy. No files have been lost.
- Four drives have been replaced since late July: three on zshare5c and one on zshare5b. The latest swap was on 11 September. The rebuild finished that night and checked out clean.
- The monthly integrity scan on 6 September read every volume end to end, and data and redundancy matched on all six. It turned up read errors on two drives in particular, and the RAID corrected every one. One of those two drives was replaced on 11 September. The other is being watched.
- Two drives are being watched (below). Both are still in service, and a spare is on the shelf for each.
- Three volumes are 100% full. Maintenance can't fix that; it needs new storage.
- No failure is known right now. One routine item is open: a replacement 4 TB spare, requested from Jeannine on 11 September.
There is some urgency, though. These are 2012 and 2013 machines, well past any normal service life. Nothing is known to be failing today, but their age and the statistics for hardware this old say we are getting steadily closer to a failure we can't see coming, and one we won't recover from. RAID6 covers individual drive failures; it does not cover the machines themselves. Replacing drives keeps the archive alive for now, but it does not make the machines younger.
Volumes
Each volume is RAID6: eight drives, any two of which can fail without losing a file.
| Server | Share | Built | Drives | Usable | Used | Status |
|---|---|---|---|---|---|---|
| zare4 | zshare4a | Aug 2012 | 8 × 3 TB | 17 TB | 100% | Healthy, 8/8 |
| zare4 | zshare4b | Aug 2012 | 8 × 3 TB | 17 TB | 95% | Healthy, 8/8 |
| zare4 | zshare4c | Apr 2015 | 8 × 6 TB | 33 TB | 100% | Healthy, 8/8 |
| zare5 | zshare5a | Jun 2013 | 8 × 4 TB | 22 TB | 93% | Healthy, 8/8 |
| zare5 | zshare5b | Jun 2013 | 8 × 4 TB | 22 TB | 85% | Healthy, 8/8 |
| zare5 | zshare5c | Apr 2017 | 8 × 10–12 TB | 55 TB | 100% | Healthy, 8/8 |
Drives being watched
| Share | Drive (label serial) | Age | Current state | Plan |
|---|---|---|---|---|
| zshare5c | ZA20QMVH | ~9 years (2017 original) | 12,016 reallocated sectors: bad spots the drive has mapped out and worked around. Up 1,072 since August, a period that included the September scan. Unchanged since 11 September. No pending sectors. | Replace if it fails. The 12 TB spare is held for it. |
| zshare4b | WD-WCAWZ2253727 | 13.6 years (2012 original) | Returned read errors during the September scan, all corrected by the RAID. No errors since. 61 pending sectors, unchanged since 7 September. | Replace if it fails. The 3 TB spare is held for it. |
Drives are replaced when they fail or show clear signs of failing, not when a counter merely rises. RAID6 plus a spare on the shelf is what makes that reasonable: a single drive failure costs a rebuild, not files.
Drive replacements since July 2026
| Date | Share | What happened |
|---|---|---|
| 24 Jul | zshare5c | A drive failed. Replaced with a 10 TB spare, and the volume rebuilt clean. |
| 29 Jul | zshare5c | A drive with over 16,000 reallocated sectors was replaced before it could fail. Replaced with a 12 TB spare, and the volume rebuilt clean. |
| 4 Aug | zshare5c | A drive that failed outright on 1 August was replaced with a 12 TB spare. The volume rebuilt clean by 5 August. |
| 11 Sep | zshare5b | A drive failing reads in normal use was replaced with the 4 TB spare. The volume rebuilt clean the same night. |
All four swaps were done by Jeannine, and each drive was identified correctly the first time.
Monitoring
- Every drive on both machines runs its own health monitoring, which emails Drew if a drive reports unreadable sectors, errors or a failed self-test. Each data drive runs a short self-test every week.
- The RAID emails Drew if a drive fails, and each machine sends a daily status email.
- The integrity scan runs on the first Sunday of each month. The next one is 4 October.
- Every drive's counters, and the RAID's own error counts, are reviewed by hand each month after the scan.
- System logs from both machines are copied off-site. The configuration needed to rebuild either machine onto a new boot drive is kept on file.
- Both machines had a security review in August and early September. Unused services were removed or switched off, and remote access was tightened.
Spares on hand
| Size | Qty | For |
|---|---|---|
| 20 TB | 1 | Earmarked for the Synology units |
| 14 TB | 1 | Earmarked for the Synology units |
| 12 TB | 1 | zshare5c |
| 8 TB | 1 | Stand-in for any 3, 4 or 6 TB position |
| 6 TB | 2 | zshare4c |
| 4 TB | 0 | zshare5a / zshare5b. The last one was used on 11 September and a replacement has been requested. Until it arrives, the 8 TB can stand in. |
| 3 TB | 1 | zshare4a / zshare4b |
A larger drive can always replace a smaller one, because the volume uses only the space it needs. zshare5c already runs two 12 TB drives in 10 TB positions.
Complete drive inventory
Every drive in both machines, measured 14 September 2026. At least 22 of the 48 are the drives their volume was built with. The rest have been replaced over the years.
- Age: how long the drive has been spinning, in years.
- Realloc: bad spots the drive has mapped out. 0 is ideal. A small number that stays put is fine. A climbing number is the warning sign.
- Pending: spots the drive suspects are bad but has not yet mapped out.
zare4
zshare4a (8 × 3 TB, built August 2012)
| Serial | Model | Age | Realloc | Pending | Note |
|---|---|---|---|---|---|
| WD-WMAWZ0265142 | WD30EZRX | 13.6 | 0 | 0 | |
| WD-WMAWZ0324492 | WD30EZRX | 13.6 | 0 | 0 | |
| WD-WMC4N0DDT38W | WD30EZRX | 4.9 | 0 | 0 | |
| WD-WCAWZ2344412 | WD30EZRX | 3.9 | 0 | 0 | |
| WD-WC81Y9051052 | WD30EZRX | 2.4 | 0 | 0 | |
| WD-WMAWZ0361626 | WD30EZRX | 13.6 | 0 | 0 | |
| WD-WMAWZ0408925 | WD30EZRX | 13.6 | 0 | 0 | A small number of read errors corrected by the RAID, date unknown. None since 11 September |
| WD-WMAWZ0342551 | WD30EZRX | 13.6 | 0 | 0 |
zshare4b (8 × 3 TB, built August 2012)
| Serial | Model | Age | Realloc | Pending | Note |
|---|---|---|---|---|---|
| WD-WMAWZ0274013 | WD30EZRX | 13.6 | 0 | 0 | |
| WD-WMAWZ0198152 | WD30EZRX | 13.6 | 0 | 0 | |
| WD-WCAWZ2261213 | WD30EZRX | 13.6 | 0 | 0 | |
| WD-WCAWZ2078646 | WD30EZRX | 2.7 | 0 | 1 | Unchanged since July |
| WD-WCAWZ2253727 | WD30EZRX | 13.6 | 0 | 61 | Watched (see above) |
| WD-WMC1T3071834 | WD30EZRX | 7.3 | 0 | 1 | New 26 September |
| WD-WMAWZ0324385 | WD30EZRX | 13.6 | 0 | 0 | A small number of read errors corrected by the RAID, date unknown. None since 11 September |
| WD-WCAWZ2261050 | WD30EZRX | 13.6 | 0 | 0 |
zshare4c (8 × 6 TB, built April 2015)
| Serial | Model | Age | Realloc | Pending |
|---|---|---|---|---|
| WD-C82DWNAK | WD60EFZX | 2.1 | 0 | 0 |
| WD-WX22D24DM29Z | WD60EFPX | 0.4 | 0 | 0 |
| WD-WX11DB4H8KNY | WD60EFRX | 2.6 | 0 | 0 |
| WD-WX11D58LZ72R | WD60EFRX | 7.0 | 0 | 0 |
| WD-WX21D48FD47L | WD60EFRX | 6.9 | 0 | 0 |
| WD-WX51D89HDFEA | WD60EFAX | 4.8 | 0 | 0 |
| WD-C80T9S3G | WD60EFZX | 4.0 | 0 | 0 |
| WD-WX11D36JRUAF | WD60EFRX | 5.5 | 0 | 0 |
zare5
zshare5a (8 × 4 TB, built June 2013)
| Serial | Model | Age | Realloc | Pending |
|---|---|---|---|---|
| ZDH0515X | ST4000DM000 | 7.9 | 0 | 0 |
| ZTT5XVJL | ST4000DM004 | 2.3 | 0 | 0 |
| W30066MR | ST4000DM000 | ~13.0* | 0 | 0 |
| ZDHAXA6M | ST4000VN008 | 3.3 | 0 | 0 |
| W3005NVH | ST4000DM000 | ~12.9* | 0 | 0 |
| S300LKHH | ST4000DM000 | 1.6 | 0 | 0 |
| ZFN5P3EV | ST4000DM004 | 0.7 | 0 | 0 |
| W3006K96 | ST4000DM000 | ~13.0* | 0 | 0 |
zshare5b (8 × 4 TB, built June 2013)
| Serial | Model | Age | Realloc | Pending | Note |
|---|---|---|---|---|---|
| ZFN1MZA8 | ST4000DM004 | 7.6 | 0 | 0 | |
| ZFN5T3N2 | ST4000DM004 | new | 0 | 0 | Installed 11 Sep 2026 |
| W3005M66 | ST4000DM000 | ~12.9* | 0 | 0 | |
| W3005QTA | ST4000DM000 | ~12.9* | 0 | 0 | |
| ZFN5RQF0 | ST4000DM004 | 0.4 | 0 | 0 | |
| WD-WCC4E6ZPE2R8 | WD40EZRZ | 9.5 | 0 | 0 | |
| W30058V3 | ST4000DM000 | ~12.9* | 0 | 0 | |
| ZFN1KWTX | ST4000DM004 | 6.8 | 0 | 0 |
\* These six drives' hour counters have reset, so their ages come from head-flying time instead.
zshare5c (8 × 10–12 TB, built April 2017)
| Serial | Model | Size | In service since | Realloc | Pending | Note |
|---|---|---|---|---|---|---|
| ZA21214K | ST10000VN0004 | 10 TB | 2017 | 8 | 0 | Unchanged since July |
| ZA210YA7 | ST10000VN0004 | 10 TB | 2017 | 8 | 0 | |
| ZA20CYVN | ST10000VN0004 | 10 TB | 24 Jul 2026 | 0 | 0 | |
| ZA20QMVH | ST10000VN0004 | 10 TB | 2017 | 12,016 | 0 | Watched (see above) |
| ZA20V0V8 | ST10000VN0004 | 10 TB | 2017 | 0 | 0 | |
| ZZ30LAEV | ST12000VN0008 | 12 TB | 4 Aug 2026 | 0 | 0 | |
| ZA20TS1E | ST10000VN0004 | 10 TB | 2017 | 0 | 0 | |
| ZZ30LASR | ST12000VN0008 | 12 TB | 29 Jul 2026 | 0 | 0 |
The 2017 drives' hour counters have reset, so they are dated from the volume's April 2017 build.
Each machine also has a 60 GB SSD boot drive. It holds only the operating system, and both report healthy.
Capacity
zshare4a, zshare4c and zshare5c are 100% full; the other three are at 85–95%. That leaves no room to stage files during maintenance and no room for new recordings on those volumes. The volumes can't be expanded in place, and putting larger drives into machines this old would take years to gain any space.
The sensible path is new storage rather than more drives for these machines: a further Synology unit with modern 20–28 TB drives, built the same way as the newer storage already on site. That is a budgeting matter for CoF, and not one to leave open-ended: nothing is known to be failing, but every month on this hardware brings an unrecoverable failure closer. Until then, these machines are kept running and drives are replaced as they fail.
Thanks to Jeannine, whose work at the rack is the reason this archive is still healthy.