NAS (any brand) external USB drive bay - issues

[sfx2000]

Wonder why the kernel is trying to mount that ext4 filesystem as ext3 - I mean it'll work, but hardly proper with very large filesystems (4TB is at the limit of ext3). That's a script somewhere inside the synology firmware that does that with hotplug..

Keep in mind that e2fsk checks the integrity of the file system, not of the files within - as i mentioned earlier, hashing out both sides and diffing the hashlists to see if there is any corruption of the files themselves would be a prudent move, esp as we know there's some random flakiness with the firmware into the jbod box with usb3

 

[sfx2000]

Switch settings inside this box are

• Normal. Two separate drives
• RAID0 striped
• RAID1 Mirror
• JBOD "concentrates 2 drives to a single drive" is says

I played with RAID0 mode and JBOD mode. Formatted both. Currently using JBOD mode since this is backup data, not critical.

Makes sense as the Prolific chip basically creates a raidset and presents it as an abstracted device if RAID0/1/linear span if told to by the switch (again, chip select drives that decision by the firmware). Doesn't seem to be a way to check the consistency of the raidset itself, at least not from a host OS level.

I suspect it's not doing uasp, but usb bulk storage

I wouldn't trust that box under linux - windows maybe, but not under linux - too many things that can go wrong there.

 

[sfx2000]

I plugged its USB3 cable into my windows 7 PC and ran the ext4 mounter program: "Disk Internals Linux Reader". It saw the volume and its name. But it wouldn't mount. I've used that windows program before, with 1-drive ext4 disks.

Yep, makes me think firmware problem with the prolific chip... bad geometry so ext4 mounter can't make sense of it.

 

[sfx2000]

Here's the data sheet on the PL chip...

 

[sfx2000]

Fun little exercise - someone was have a similar problem over in the AsusWRT-RMerlin sub-forum, now we know why, and why it's not wise to use one of these boxes with an embedded linux platform.

 

[stevech]

If there was a sector address scramble (LBA mapping to two drives) error in the chip firmware, I'd guess that it would affect inodes/directories as well as blocks containing data, and fsck would find errors. Yes?

 

[stevech]

another thought... there are 100s of GB of drive images produced by Acronis that were copied to the box. Acronis can validate the structure of that. Also the zip files can be validated.

 

[sfx2000]

With ext4 and the journal, maybe e2fsk could tell, but again, it's not checking the files against the source, just that the files are there in the trees/nodes

The Acronis check, that is a excellent idea, as it has a concept of what is in the image by the metadata in the image file itself.

Same with the zip archives..

 

[stevech]

another curiosity....

• Start with no USB drives plugged in.
• Plug in my Seagate USB3 2TB drive (2.5in. bus powered). Comes up normally in the NAS external devices display.
• Plug in the 2 bay enclosure. Does not come up in the NAS. (Hot-plug does work after reboot)
  
• Leave both plugged in.
• Reboot NAS.

Now both are visible as shown in the attached screen grab. No hot plug ... discovery during boot.

Blame Linux kernel? Or blame Synology?

 

[sfx2000]

blame Synology

'nuff said...

 

[stevech]

Synology just dismissed this report from me... "We don't support DAS". I counterered, but to no avail.
Oh well. The Cheapie enclosure for JBOD/RAID0 is what it is.

 

[sfx2000]

Oh well. The Cheapie enclosure for JBOD/RAID0 is what it is.

Some things I would buy from Shenzen - things that might do me harm... perhaps not...

 

[sfx2000]

Synology just dismissed this report from me... "We don't support DAS". I counterered, but to no avail.

After stepping thru this thread - I agree with them...