The article explains about the ATA errors seen on console messages.
The symptoms are
Below error logs can be seen from console:<<<
(ada1:ata0:0:1:0): Error 5, Retries exhausted g_vfs_done():ada1p2[WRITE(offset=140206080, length=4096)]error = 5 (ada1:ata0:0:1:0): WRITE_DMA. ACB: ca 00 da 85 13 40 00 00 00 00 08 00 (ada1:ata0:0:1:0): CAM status: ATA Status Error (ada1:ata0:0:1:0): ATA status: 41 (DRDY ERR), error: 04 (ABRT ) (ada1:ata0:0:1:0): RES: 41 04 da 85 13 00 00 00 00 08 00 (ada1:ata0:0:1:0): Retrying command (ada1:ata0:0:1:0): WRITE_DMA. ACB: ca 00 da 85 13 40 00 00 00 00 08 00 (ada1:ata0:0:1:0): CAM status: ATA Status Error (ada1:ata0:0:1:0): ATA status: 41 (DRDY ERR), error: 04 (ABRT ) (ada1:ata0:0:1:0): RES: 41 04 da 85 13 00 00 00 00 08 00 (ada1:ata0:0:1:0): Retrying command
<snip…>
<<<
AdaX is the virtual disk only present in qfx-tvp models. When these above symptoms are seen, the only way is to power cycle the device. These error mean that the SSD went into error state causing disk access failures. These can be transient hardware ssd errors that recovers after a reboot of the device. Sometimes there can be fatal errors that needs the device to be RMA'd.
In such cases, host logs could show below errors.
ata6: hard resetting linkata6.00: disabledata6.00: device reported invalid CHS sectorata6: EH completesd 5:0:0:0: [sdb] Unhandled error code
In some case there are no errors seen in the hostlogs. This might be a Qemu Disk access errors. QFX platforms are running Junos as quest in Linux host using Quick Emulator. If quick emulator block io module gets into error state, further it may not be able to provide disk access provisioning to guest. This can freeze the guest disk io operations.Fixes are built on later images to automatically reboot the device after encountering such issue.
2023-04-24 version 12023-05-16 version 2
2025-04-14 modified the content and font for external view