Help to understand an error message on hdd

[StefanoN]

Hello
After an update of Skynet with an error (I have seen the error go by quickly and it seems to me that it indicates filesystem in read only mode) I had to reboot my router rt-ac86u.
After the reboot on the system log I see these messages:
Jun 4 20:59:42 kernel: sd 0:0:0:0: [sda] FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
Jun 4 20:59:42 kernel: sd 0:0:0:0: [sda] Sense Key : Medium Error [current]
Jun 4 20:59:42 kernel: sd 0:0:0:0: [sda] Add. Sense: Unrecovered read error
Jun 4 20:59:42 kernel: sd 0:0:0:0: [sda] CDB: Read(10) 28 00 04 40 00 28 00 00 20 00
Jun 4 20:59:42 kernel: blk_update_request: critical medium error, dev sda, sector 71303208
Jun 4 20:59:42 amtm disk-check: Disk check done on /dev/sda1
Jun 4 20:59:42 hotplug: USB ext4 fs at /dev/sda1 mounted on /tmp/mnt/SSD
Jun 4 20:59:42 usb: USB ext4 fs at /dev/sda1 mounted on /tmp/mnt/SSD.

The router is ok, but the messages seem to indicate a disk failure. Is it correct ? The hdd is a SSD, the router firmware is the 314.18

Thanks to all

 

[ColinTaylor]

Yes the error indicates that the router couldn't read sector 71303208 which usually indicates a failing HDD. It's a bit odd to see that with an SSD though. Perhaps the SSD was in sleep mode?

P.S. Check the cable between the router and SSD is secured properly.

 

[StefanoN]

Yes the error indicates that the router couldn't read sector 71303208 which usually indicates a failing HDD. It's a bit odd to see that with an SSD though. Perhaps the SSD was in sleep mode?

Thanks for the answer. The Enable HDD hibernation is set to NO.

To replace the SSD I should remove Skynet, entaware, swap file, and reinstall all on a new disk ?
The others tools as Diversion and Yazfi ... seem installed on jffs/script

Thanks

 

[OzarkEdge]

FWIW, I got tired of seeing odd behavior from a USB HDD attached to my 86U. Issues included:
o not knowing if it suffers during a router reboot or disruption;
o random file system errors found when scanned by the router tool;
o hobbled by USB 2.0 Mode since 3.0 Mode interferes with 2.4 GHz WLAN client stability;
o not 100% compatible with Windows Explore usage;
o requires using deprecated Windows SAMBA 1.0 client;
o 'unexpected network errors' in backup software that otherwise has always performed rock solid locally and over the network.

So, I moved my USB HDD back to a local machine where it performs like it should without weirdness... you can't have storage misbehaving for indeterminate reasons.

OE

 

[Josh Schneider]

After two years of stable use, I just had an SSD get corrupted while attached to my RT-AC68P (USB3 port). I have another SSD and will be getting the old one replaced under warranty.

Two questions, I guess, assuming I'm going to keep doing this:

1) Should I leave "Enable HDD Hibernation" set to "No" or switch it to "Yes"?
2) Am I somehow shortening the drive's life by using it this way, or does anyone know? I'm pretty much using it for backup and some streaming media, so not much writing, and it mostly sits idle much of the time

Thanks

 

[ColinTaylor]

Define "corrupted". The filesystems on storage devices often become corrupted, typically after unexpected power loss. That's not a reason to replace the drive though as it's just a soft error that reformatting the drive will resolve.

Hibernation should be mostly irrelevant for SSDs because unlike HDDs there are no mechanical parts. So I can't think of a reason to enable it.

Drive life should not be an issue. These drives are intended for day to day use in PCs where the read/write activity is much much higher.

 

[Josh Schneider]

@ColinTaylor, unrecoverable bad superblock errors with fsck, then ran smartctl and the drive reported failure and had gone into RO mode. I have an RMA ticket for it. It's a 2-year-old HP NVME drive, EX920.

Thank you for the additional details. I think in this case I got a lemon. The drive was recommended for reliability and lifespan, among other things, and I am not using it intensively.

The router may not manage it's power so well, not sure. It's enclosed in a MyDigitalSSD M2X with a JMicron controller, and when I pulled it, it was quite cool and looking fine. I can still access the drive in Windows, but no longer in Linux. When it started having issues, I could not write to it from a file manager, but I could still see its contents with it mapped through SMB (Windows, Linux, Android, whatever).

 

[ColinTaylor]

@Josh Schneider Probably just bad luck as you say. The only thing that comes to mind that might cause that kind of fault would be some sort of electrical spike. That seems unlikely though as any spikes from the mains would have to get through both the router and the M2X. The lack of any earth connection to the router might be an issue with static charges. I'm just thinking aloud here.

 

[Josh Schneider]

Thanks, @ColinTaylor. Yeah, I don't think an electrical spike is likely, since the router and the M2X also seem fine, and nothing happened in the house prior to the issue cropping up. I also have whole-house surge protection on my panel.

 

[L&LD]

I don't think that the drive life of an SSD connected to a router or a computer is equivalent? On a computer, the OS will be using TRIM and everything else it has available to keep it healthy.

On the router, it is just the SSD's firmware that will decide how healthy it will stay. For many drives, that may be 'not long'. Or, certainly not as long as it would be when connected to a supporting OS.

 

[Josh Schneider]

I now am not sure the problem isn't with the JMicron enclosure I've been using. However, I think @L&LD's point is worth understanding. TBD where my issue is. I don't have a computer that can take this drive natively. I've ordered another enclosure, and have an RMA if I need it.

 

[thecheapseats]

I now am not sure the problem isn't with the JMicron enclosure I've been using.

two things - (1) long experience here avoiding any chipsets (jmicron particularly) other than asmedia on any ext-disk linux or embedded linux device enclosure for the very corruption symptoms you mentioned...

(2) per your comment, without a way to natively test - using a diskpart 'clean' command (on win) or desktop linux-dist equivalent utility with a known un-weird chipset enclosure may well get that nvme device functioning again real fast... been there...

additionally, there have been numerous discussions here regarding ssd-trim on linux as well as hanging ssds on these routers, testing for trim functioning and 'does it matter?' - if you're so inclined to spend a lot of time reading... I wasn't...

 

[Josh Schneider]

Thanks, @thecheapseats.

I think the new unit I'm getting (Plugable) has a Realtek RTL9210. ALthough people are saying firmware updates may have fixed issues with some JMicron chipsets, I figure it's worth trying another, and I chose this over an ASMedia-driven unit because of some other reviews. If you have any thoughts on that, I'd like to know.

I guess I'm going to try to read up on options for my local NAS-like setup. I just read an article about using a Raspberry Pi to create cheap NAS with RAID, and I expect something like that might be configured to protect the drive's lifespan. I'll search around for further discussions here, since my original thought was that my Merlin setup would provide me with basically a cheap NAS option.

 

[thecheapseats]

I've was told firsthand by a friend that went that way - Plugable - that their support will work with customers to make their devices work, as there have been issues with the realtek - but you pay a serious premium for their hardware in general...

being a cheap guy but testing to confirm what consistently works, doesn't make me jump thru hoops or puke data on my shoes, I only run asmedia devices... but that's me, and only after being advised by pcb designers I've known for years who touch this stuff everyday and have their shortlists of preferred IC devices...

What must be considered when hanging ssd devices off of any usb port (whether it's a desk/lap-computer, router, a raspi or a nas) - is first, the usb chipset on the device - a leap of faith and usually mystery meat (unless you take it apart and actually 'look' at the pcb) - and next is the external device's chipset(s) that hosts the ssd or nvme... as well, consider the supported xfr protocols...

disclaimer: I hate usb for data xfr that's important - it's a hot mess moving target, often implemented incorrectly and I never rely on it if I don't have to... it's a toy that's fine for a mice or keybds...

do yourself a favor if budget permits - get a real nas... someday the new pi compute module4 with a pcie card would be miles better than anything usb...

best info on raspi4 and usb/ssd is here (dated as to booting pi4s over usb) - but the hardware info is priceless:

Raspberry Pi 4 USB Boot Config Guide for SSD / Flash Drives

The Raspberry Pi 4 is finally here and has a lot of exciting changes. One very major downside is that it doesn’t support true USB booting yet out of the box (like the 3 series did).

jamesachambers.com

 

[Josh Schneider]

Thanks, @thecheapseats. Will take this all into consideration. I may just try out the Plugable to see if my drive isn't actually hosed. But I've seen ASMedia options, as well. I'm getting a "used" (warehouse) enclosure, so it's not ridiculously expensive. If it works, I may just keep it. I have to see what's up with firmware updates for that chipset now.

I appreciate the advice about setting up a NAS and issues relying on USB. Will just have to see what kind of effort I'm willing to put into figuring out a good alternative option. I'm also still looking for a cloud option to backup from local, and I think that's crucial to really avoid issues down the road. Don't have that figured out just yet.

 

[thecheapseats]

also an afterthought - how were you powering that external nvme enclosure.?... relying upon usb Power may have bit you if the power xfr on the usb bus (voltage regs, caps, etc.) went south, as you mentioned it was fuctioning for years... an issue to be considered...

 

[thecheapseats]

<snip> may just try out the Plugable to see if my drive isn't actually hosed.<snip>

diskpart on a win machine (if you run windows) is your friend if you get that far..

 

[Josh Schneider]

@thecheapseats, thanks. My options really are to try out another enclosure or find a computer that can host the drive natively. My wife has a loaner from the school where she teaches, but it is locked from booting via a usb stick. If I could change that, I'd be able to test the drive natively. As it is, I need to try out my Plugable enclosure.

I think I'm out of luck with the M2X, since it's out of warranty, but we'll see. I managed to update its firmware, so it at least still communicates with my laptop via USB.

 

[thecheapseats]

if you plug that nvme in the m2x enclosure into a win laptop, does diskpart 'see' it on the command line?...

be careful if you've never used diskpart... way too easy to wipe a boot drive with it if you're not careful...

if it 'lists' (using list disk) - 'select' it (see the man page or help), then use the 'clean' command... it may well enable it to be seen in win's idiot 'disk mangement gui' where it can be initialized and then formatted...

 

[thecheapseats]

nvme enclosures for usb are rare ducks - not widely used/needed so finding one that works is tough... I'm not a ugreen fan (chinese crap) but I found this one with an asmedia chipset in my junk box to occasionally 'clean' any nvme that's lost it's damn mind...

https://www.amazon.com/dp/B07NPFV21K/?tag=snbforums-20

for isolating power problems, I did wire a breakout for the usb-C connector that separates the power from the data lines to eliminate flaky power issues from an upstream usb device - but even without that, it may work in a pinch to do a diskpart 'clean'...