SSD CACHE ERROR

[Jimbuoy]

A few days ago my TS932X gave some errors regarding the SSD drives we use as a Cache

Message: [Storage & Snapshots] SSD cache RAID group "2" is in degraded mode.

Message: [Hardware Status] "Host: SSD 4": Disconnected.

Message: [Storage & Snapshots] Changed SSD cache type from "Read-Write" to "Read-Only".


I powered down and took SSD 4 out and reseated it .. put it back in as I found it hard to believe that it had failed .. only 10 months old. Rebooted and in Storage Disc health etc it reports the disc as being good.

But the cache type is now read only. How do I get back to read and write.. should I really have replaced teh SSD?

Thanks

Jim

[storageman]

Are they approved?
Have you SMART checked them?
So if it says read only, is it saying cache RAID is degraded?

You could delete the cache and recreate it but only if the SMART check shows no errors. (Make sure no IO going on)

[Jimbuoy]

Yes they are approved. I've run the rapid test on them all and they are all good. I'll try deleting the cache and recreating it.

[dolbyman]

the 932X has 4x2.5 inch SSD bays, the only reason to power down the NAS would be NVMe storage, everything else is done HOT, never shutdown the NAS to swap drives

[Jimbuoy]

Thanks for that Dolbyman

[Jimbuoy]

Re Creating the Cache has given me back Read-Write.

[aknauf]

I have a similar issue. I rebooted my QNAP TS-932X NAS prior to upgrading the firmware and when it came up it detected only one of my 2 SSDs - although the 5 HDDs were fine. I removed the SSD and re-inserted it, but it still failed. tried it in the other bays, still failed to detect it. Tried a different SSD, still failed to detect it. Tried all of my various SSDs (I have 3) in all of the 4 SSD bays - it wouldn't detect any of them, although with some a voice would say "unrecognised drive, please replace it." Several reboots later, still no SSDs being detected.

Eventually I shut the NAS down completely and disconnected the power cable, then on the next startup everything was back to normal!

Throughout this time I would get dmesg output similar to the following:

Code: Select all

[  981.284017] ata6: exception Emask 0x10 SAct 0x0 SErr 0x4000000 action 0xe frozen
[  981.291420] ata6: irq_stat 0x00000040, connection status changed
[  981.297426] ata6: SError: { DevExch }
[  981.301097] ata6: hard resetting link
[  983.300088] ata6.00: WARNING: PHY hardreset for port repeated 100 times
[  987.920107] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  992.920091] ata6.00: qc timeout (cmd 0xec)
[  992.924186] ata6.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[  992.930282] host 0x1c36:0x0031. port_no 1:1.
[  992.934548] ata6: hard resetting link
[  992.938226] ata6.00: SSS WA failed after 10 usec, sstatus = 0x123
[  993.060091] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 1003.060094] ata6.00: qc timeout (cmd 0xec)
[ 1003.064191] ata6.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 1003.070285] host 0x1c36:0x0031. port_no 1:1.
[ 1003.074550] ata6: limiting SATA link speed to 1.5 Gbps
[ 1003.079682] ata6: hard resetting link
[ 1003.083364] ata6.00: SSS WA failed after 10 usec, sstatus = 0x123
[ 1003.200093] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 1033.200092] ata6.00: qc timeout (cmd 0xec)
[ 1033.204189] ata6.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 1033.210283] host 0x1c36:0x0031. port_no 1:1.
[ 1033.214549] ata6: hard resetting link
[ 1033.218227] ata6.00: SSS WA failed after 10 usec, sstatus = 0x123
[ 1033.340088] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 1033.346265] ata6: EH complete

It might also be interesting to note that eventually this dmesg output stops, too. As in it would normally spit something out when I insert a drive, but then it would stop after a while of reinserting drives.

Anyone else had similar a experience?


ADK

[inquisitor]

A hint for those who like myself stumble upon this thread after receiving the surprising error message "Changed SSD cache type from 'Read-Write' to 'Read-Only'.":

Apparently QTS automatically switches from Read-Write to Read-Only if the Temperature Alarm threshold set manually by the user (under Storage & Snapshots > Storage > Disks/VJBOD > SSD > Disk Health > Settings > Temperature Alarm) is exceeded. This is an unexpected and undocumented behaviour as the menu just says "enable temperature alarm for this drive" but nothing about configuration changes.
Worse, it looks like there is no way to switch back to Read-Write other than entirely removing the SSD from the Cache Acceleration menu and reconfiguring it with "Read-Write" set.