Description

kernel: ata6.00: failed command: WRITE DMA logs are seen in host logs

Symptoms

Device is not accessible via management or console.

After recovering the device, looking at host logs we see below messages.

Write DMA failed.

2023-04-14T14:46:39.143696+00:00 jtac-qfx5100 kernel: ata6.00: failed command: WRITE DMA
2023-04-14T14:46:39+00:00 dcpfe0 kernel: ata6.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
2023-04-14T14:46:39.181144+00:00 jtac-qfx5100 kernel: ata6.00: cmd ca/00:18:58:54:dd/00:00:00:00:00/e1 tag 14 dma 12288 out
2023-04-14T14:46:39.181162+00:00 jtac-qfx5100 kernel:   res 40/00:01:00:4f:dd/00:00:00:00:00/00 Emask 0x4 (timeout)
2023-04-14T14:46:39.327854+00:00 jtac-qfx5100 kernel: ata6.00: status: { DRDY }

It tries to hard reset the link but it fails.

2023-04-14T14:46:39+00:00 dcpfe0 kernel: ata6: hard resetting link
2023-04-14T14:46:49.437423+00:00 jtac-qfx5100 kernel: ata6: hard resetting link
2023-04-14T14:46:49+00:00 dcpfe0 kernel: ata6: COMRESET failed (errno=-16)
 
After few retires, it give up and the link is shut down.

2023-04-14T14:47:34+00:00 dcpfe0 kernel: ata6: COMRESET failed (errno=-16)
2023-04-14T14:47:34+00:00 dcpfe0 kernel: ata6: limiting SATA link speed to 1.5 Gbps
2023-04-14T14:47:36.093850+00:00 jtac-qfx5100 vehostd[5864]: vjunos_vmm_periodic: [vjunos0] vehostd missing keepalive
2023-04-14T14:47:39+00:00 dcpfe0 kernel: ata6: COMRESET failed (errno=-16)
2023-04-14T14:47:39+00:00 dcpfe0 kernel: ata6: reset failed, giving up
2023-04-14T14:47:39.711585+00:00 jtac-qfx5100 kernel: ata6.00: disabled

 

Solution

To recover SSD at present power cycle is required. Reboot will not resolve the issue.

Modification History

04-21-2023 initial KB.