Article Detail

Description

This article helps you understand and troubleshoot QFX5K/EX4600 series switches' inaccessibility with multiple port failures caused by READ_DMA errors.

Symptoms

1) You might lose in-band access to the device.

2) The following logs will be observed on the Console:

ad0: FAILURE - READ_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=125695

ad0: FAILURE - READ_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=125695

ad0: FAILURE - READ_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=125695

ad0: FAILURE - READ_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=125695

ad0: FAILURE - READ_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=125695

ad1: FAILURE - WRITE_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=8137093

ad1: FAILURE - WRITE_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=79

ad1: FAILURE - WRITE_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=159

3) Multiple ports might go down, leading to a service outage.

4) The following commit error might also be seen along with the previous logs:

error: /config/juniper.conf+: write failed: Broken pipe
error: /config/juniper.conf+: was not properly closed: Broken pipe
error: commit failed: commit to file failed during export to /config/juniper.conf+

Solution

The following error logs "FAILURE - READ_DMA status=41<READY,ERROR> error=4<ABORTED>" signifies that the switch is unable to read from the disk, thus indicating possible disk failure. It could be either a fatal or a transient error in the SSD.

1) As a recovery, perform power-cycle of the switch which could resolve the transient SSD error.

2) If the issue persists following a power-cycle, this implies a hardware fault, and therefore, replacement of the switch is required.

Modification History

2025-03-25 : Article Created

2025-05-12 : Added EX4600s to Categories.

2025-8-13 :

Added the following log examples:

ad1: FAILURE - WRITE_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=8137093
ad1: FAILURE - WRITE_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=79
ad1: FAILURE - WRITE_DMA status=41<READY,ERROR> error=4<ABORTED> LBA=159

Added the commit error based on the behavior seen in case 2025-0812-813273