在向ZFS文件系统复制大量数据时，存储空间损坏

kiko_

设备环境：物理机 N100 8G 14T SATA单盘 ZFS文件系统 通过m.2转6SATA卡转接 ZFS ARC设置4G

BUG现象：从 外接存储 NTFS 硬盘向 ZFS存储空间复制10T的数据，小文件较多，都是短视频的拍摄素材

某个时间点ZFS文件系统提示降级，存储空间损损坏。但还在正常复制文件。

随即停止了文件复制

root@:~# zpool status -x
  pool: trim_c956f230-7561-4a89-a81f-0d926429e355
 state: DEGRADED
status: One or more devices has experienced an unrecoverable error.  An
        attempt was made to correct the error.  Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
        using 'zpool clear' or replace the device with 'zpool replace'.
   see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-9P
config:

        NAME                                       STATE     READ WRITE CKSUM
        trim_c956f230-7561-4a89-a81f-0d926429e355  DEGRADED     0     0     0
          1653e609-c246-402f-95cf-16e746ca354f     DEGRADED     0    40     0  too many errors

errors: No known data errors

联系方式：17604587671 481群 - N150

日志文件：飞牛分享【Debug_Log_20251202195156.zip】，点击链接下载文件，App打开可转存到NAS：https://share.fnnas.net/s/7593fcdf9f4746529c

附件过大无法上传可以通过飞牛外链分享或者百度网盘提供日志文件

飞牛技术同学

[Tue Dec  2 19:08:24 2025] ata4.00: irq_stat 0x80400040, connection status changed
[Tue Dec  2 19:08:24 2025] ata4: SError: { PHYRdyChg DevExch }
[Tue Dec  2 19:08:24 2025] ata4.00: failed command: WRITE FPDMA QUEUED
[Tue Dec  2 19:08:24 2025] ata4.00: cmd 61/08:70:88:fe:00/00:00:28:01:00/40 tag 14 ncq dma 4096 out
                                    res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x10 (ATA bus error)
[Tue Dec  2 19:08:24 2025] ata4.00: status: { DRDY }
[Tue Dec  2 19:08:24 2025] ata4.00: failed command: WRITE FPDMA QUEUED
[Tue Dec  2 19:08:24 2025] ata4.00: cmd 61/50:78:70:09:01/04:00:28:01:00/40 tag 15 ncq dma 565248 out
                                    res 40/00:01:00:00:00/00:00:00:00:00/00 Emask 0x10 (ATA bus error)
[Tue Dec  2 19:08:24 2025] ata4.00: status: { DRDY }
[Tue Dec  2 19:08:24 2025] ata4: hard resetting link
[Tue Dec  2 19:08:24 2025] ata4: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[Tue Dec  2 19:08:25 2025] ata4.00: configured for UDMA/133
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#14 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#14 Sense Key : Illegal Request [current]
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#14 Add. Sense: Unaligned write command
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#14 CDB: Write(16) 8a 00 00 00 00 01 28 00 fe 88 00 00 00 08 00 00
[Tue Dec  2 19:08:25 2025] I/O error, dev sda, sector 4966121096 op 0x1WRITE) flags 0x100 phys_seg 1 prio class 2
[Tue Dec  2 19:08:25 2025] zio pool=trim_c956f230-7561-4a89-a81f-0d926429e355 vdev=/dev/disk/by-partuuid/1653e609-c246-402f-95cf-16e746ca354f error=5 type=2 offset=2542652952576 size=4096 flags=3145856
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#15 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#15 Sense Key : Illegal Request [current]
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#15 Add. Sense: Unaligned write command
[Tue Dec  2 19:08:25 2025] sd 3:0:0:0: [sda] tag#15 CDB: Write(16) 8a 00 00 00 00 01 28 01 09 70 00 00 04 50 00 00
[Tue Dec  2 19:08:25 2025] I/O error, dev sda, sector 4966123888 op 0x1WRITE) flags 0x100 phys_seg 93 prio class 2
[Tue Dec  2 19:08:25 2025] zio pool=trim_c956f230-7561-4a89-a81f-0d926429e355 vdev=/dev/disk/by-partuuid/1653e609-c246-402f-95cf-16e746ca354f error=5 type=2 offset=2542654382080 size=565248 flags=2148533376
[Tue Dec  2 19:08:25 2025] ata4: EH complete

需要排查这块14t的硬盘的硬件问题，日志出现了对应的io报错

kiko_

飞牛技术同学 发表于 2025-12-9 16:16
[Tue Dec  2 19:08:24 2025] ata4.00: irq_stat 0x80400040, connection status changed
[Tue Dec  2 19:08 ...

是的，我在向其大量复制文件，这种情况经常发生？我使用 zpool clear 修复了它

人可广大

你这明显硬件问题 ，我们zfs 没这个问题，不建议用m.2转6SATA卡转接这种转接，要想稳定上直通卡

kiko_

人可广大 发表于 2025-12-10 10:08
你这明显硬件问题 ，我们zfs 没这个问题，不建议用m.2转6SATA卡转接这种转接，要想稳定上直通卡
...

换了大电源，感觉是电源不够了

qianlongzt

kiko_ 发表于 2025-12-9 17:08
是的，我在向其大量复制文件，这种情况经常发生？我使用 zpool clear 修复了它 ...

zpool clear 不是修复，是无视这个错误。要仔细检查。

kiko_

qianlongzt 发表于 2026-1-4 12:48
zpool clear 不是修复，是无视这个错误。要仔细检查。

是的，我直接清除了，先保证业务可用，后续换了电源，好一些了。