Disk Not Initialized | LSI RAID 5 | DIY NAS | No Clue

No data loss occurred - I just want to understand what (tf) happened…

Hi!

  • I have a DIY NAS: 4790k, Asus VII Impact, 2x8GB, LSI 9361-8i, 4x HSGT HDN728080ALE604, 2x HGST HUS728T8TALE6L4, SF600, Node 304, Windows 10 Ent. This one was set up several months ago without any issues whatsoever. The LSI raid controller is set up in RAID 5.
  • Using SMB (Multichannel) via Netgear S350 switch between my Desktop and the DIY NAS, I get about ~225MB/s max. Since both Desktop and DIY NAS have only 1Gbe NICs, I added a 1Gbe USB3.0-to-Ethernet adapter each.
  • I have a backup of all the data in question, a separate Synology NAS. This one has 2x 1Gbe ports, going through the same Netgear switch.

Basically, I was copying large amounts of data from the Synology NAS to the DIY NAS, basically for the past 5-6 days, again without any issues, UNTIL I logged back into the DIY NAS to find drive D: was gone. Disk Management shows “Not Initialized”.

I have both LSI Storage Authority (Browser) and MegaRAID Storage Manager (Windows) installed. Controller is fine, no error in the log, temp is fine (Noctua 40mm cooler), RAID Array is fine, all the disk are there etc.

I checked with CrystalDiskInfo, all drives are normal temp, no warnings, no errors. I am running an extended self-test on each drive with Hard Disk Sentinel right now, ~75% done each so far no errors. Planning to do a read-write-read test afterwards.

I did not see anything in particular in the Windows logs, but I may very well have missed something there.

I am just trying to understand wtf happened. My thoughts going from SMB multichannel, to the USB-to-ethernet adapter, to dark magic.

Any thoughts much appreciated!
k

Connections have a limited life span, to prevent processes hogging all resources and thus systems becoming unresponsive. This is the TTL value (time-to-live) seen in config files. When the TTL timer runs out, the OS resets the connection (i.e. disconnects the client/server connection) and any data still flowing stops. Period, no parole/grace. Usually, the OS resets the TTL timer every time a data packet request comes in, but it appears your data flow was such that the TTL timer could run out before the data packet was actually send.

Eerrrr, stuff that, I just realized your problem is a missing drive. That I have no clue about :frowning: :raised_hands:t2:

I would understand data loss, data errors, disk errors, controller gone, etc, but it’s just beyond me how the volume in Windows is gone, when everything else seems fine …
It was a massive amount of data >2TB, but the whole volume gone is still odd.

I am thinking about switching to Unraid or just get a larger Synology DS2419+ with 2x RAID5, rather than this DIY Windows NAS …

Yeah, Win-OS is fine for the limited use-cases it was designed for, everything else, not so much :stuck_out_tongue:

Anyway, try a Linux Live-CD to see what it can find. Your missing drive may show up again, allowing some data recovery if you must, or at least you’ve established it’s a Win-10 issue.

1 Like

It’s not a good idea to run raid5 on such big drives, probability for a drive failure is very high. Should run zfs or the like if possible. If a drive fails, it’s above 50% probability for a drive failure during rebuild.

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.