CAM status: Uncorrectable parity/CRC error

Aknot

Aknot

I have had these errors in the log for a few days (new hard drives are already sent from our supplier):

Code: 
(ada3:ahcich3:0:0:0): WRITE_FPDMA_QUEUED. ACB: 61 00 98 a5 01 40 3a 00 00 06 00 00
(ada3:ahcich3:0:0:0): CAM status: Uncorrectable parity/CRC error
(ada3:ahcich3:0:0:0): Retrying command, 3 more tries remain

We have checked the data on the drive (ZFS Mirror) with zpool status which didn't show any errors, until today:

Code: 
pool: mirr_slave
state: ONLINE
status: One or more devices are faulted in response to persistent errors.
        Sufficient replicas exist for the pool to continue functioning in a
        degraded state.
action: Replace the faulted device, or use 'zpool clear' to mark the device
        repaired.
config:

        NAME        STATE     READ WRITE CKSUM
        mirr_slave  ONLINE       0     0     0
          mirror-0  ONLINE       0     0     0
            ada3    ONLINE       0     0     0
            ada2    FAULTED      3 1.15K     0  too many errors

errors: No known data errors

What are your thougths on this?
Hardware failure?
Also, the log shows ada3 and zpool status shows ada2?

Thanks
 
That's usually some kind of hardware issue,

CPU / RAM (run memtest86+ for ~6-12h to verify)
Faulty port/cable (replace cable and/or switch port)
Dying / Overheating / crappy AHCI controller (unlikely if it's Intel on-board), some early AMD chipsets can be a bit dodgy.
Dying HDD (run a long SMART test and look at the log)
 
diizzy said:
That's usually some kind of hardware issue...
Click to expand...
We have already replaced the cables with 6Gb/sec approved shielded cables, just to be sure. No difference.

I could try to bypass the hot swap bay, if there are some electrical issues in the connectors.

I have read about the RAM as a cause, but would that cause errors on the same drive all the time?

SMARTD is installed, doesn't show anything that looks suspicious.

Does a long test doing a fsck?
 
How many hours on the drive?
I replace customer drives at 26,000 hours which is 3 years of 24x7 operation.

Best done on my schedule than on the drives schedule when it decides to die.
 
bgavin said:
How many hours on the drive?
...
Click to expand...

This is what I have:

Code: 
/dev/ada2 Power_On_Hours
0x0032   020   020   000    Old_age   Always - 58679

/dev/ada3 Power_On_Hours
0x0032   020   020   000    Old_age   Always - 58847

=== START OF INFORMATION SECTION ===
Model Family:     Western Digital Red (CMR)
Device Model:     WDC WD10EFRX-68FYTN0
Serial Number:    WD-WCC4J6XXFK5P
LU WWN Device Id: 5 0014ee 20b8eff15
Firmware Version: 82.00A82
User Capacity:    1,000,204,886,016 bytes [1.00 TB]
Sector Sizes:     512 bytes logical, 4096 bytes physical
Rotation Rate:    5400 rpm
Device is:        In smartctl database 7.5/5706
ATA Version is:   ACS-2 (minor revision not indicated)
SATA Version is:  SATA 3.0, 6.0 Gb/s (current: 6.0 Gb/s)
Local Time is:    Sun Aug  9 17:27:02 2026 CEST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

I'm going to replace them both with two brand new Western Digital 7200 RPMs, before next step, starting to test NVMe's on PCIe expansions cards instead.
 
Aknot said:
We have already replaced the cables with 6Gb/sec approved shielded cables, just to be sure. No difference.

I could try to bypass the hot swap bay, if there are some electrical issues in the connectors.

I have read about the RAM as a cause, but would that cause errors on the same drive all the time?

SMARTD is installed, doesn't show anything that looks suspicious.

Does a long test doing a fsck?
Click to expand...
Hot swap bays can certainly cause issues
Memory issues can show itself in many odd ways
 
If you can rule out everything in the physical connection path (different SATA port, different cable, different enclosure slot or direct connection), then my guess would be either the disk's connector or the disk's own controller starting to give up.
 
Western Digital has a bootable WD Tools that is OS agnostic.

Start with the quick tests then work up to the long tests. It works on sectors and not file system data so it can do a sector scan on any OS.
 
I'm deeply troubled by the ada2/ada3 mixup. A problem with two sides of a mirror gets my immediate and undivided attention! I can't explain what's going on, but would want to resolve that (with caution) before anything else. I would, if I could, add a third mirror before taking any other action.

I had a similar issue CRC very recently. The system had been powered down in a seaside location for a few months. When I rebooted, ZFS took a drive offline with too many Uncorrectable parity/CRC errors. The smart report on the dead drive had no indication of any other problems. Self tests don't readily identify external problems. It was a Seagate 4TB Enterprise drive, the newest in the box.

I slotted a new drive into a spare hot-swap bay and re-silvered it into the mirror. I then shut the system down, opened the case, removed the dead drive, replaced it with the new one, and exercised the system. No problems. So no enduring problem with power, controller, or cables evident.

I installed the dead disk into a hot swap bay, wrote to 100% of the sectors and read them all back. No problems. So the disk seemed OK. I then exercised it. Still no problems. It's gone back into the spares pool with a history tag.

I put it down to salt air and a poor SATA connection -- the easiest of all disk problems to fix.

A quality pencil eraser is great for cleaning gold contacts.
 
Aknot said:
Does a long test doing a fsck?
Click to expand...
No. The drive itself does it, and it should not know about file systems.
I have had drives before, which when they were going down were picked up on by ZFS first. SMART report said nothing, then the long test came in with problems.
 
Parity/CRC error is a communication problem between the drive (or more accurately its internal controller chip) and the computer. It does not indicate that the data on the platter or the heads used to read/write the platter have a problem. Although it also does not indicate that such problems might not also exist; there could even be a common root cause that is causing multiple issues.
 
You must log in or register to reply here.
Back
Top