Network Link Fails With "mlx4_core Internal error detected" entries in the /var/log/messages Log File (Doc ID 1515524.1)

Last updated on DECEMBER 06, 2021

Applies to:

Symptoms

During normal operation, the server experienced a link going down. Inspecting the /var/log/messages file, we see the entries:

Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: Internal error detected:
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[00]: 003277db
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[01]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[02]: 20071fc2
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[03]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[04]: 003277a8
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[05]: 00000044
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[06]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[07]: d0013481
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[08]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[09]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[0a]: 00b00190
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[0b]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[0c]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[0d]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[0e]: 00000000
Dec 5 04:46:17 localhost kernel: mlx4_core 0000:19:00.0: buf[0f]: 00000000
Dec 5 04:46:17 localhost kernel: RDS/IB: rds_ib_setup_qp failed (-95)

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.