Failed to Verify the device_ip Field in the Rank Table File EI0014)

Symptom

The CANN log contains the keyword "the IP address(***) in the ranktable is inconsistent with the IP(***)address of the network adapter," as shown below:

[ERROR] HCCP(166192,eExecutor):2025-01-21-16:59:39.962.565 [ra_host.c:480]tid:167056,ra_rdev_init_check_ip(480) : [check][ip]fail, ret(129) the IP address(127.10.0.0) in the ranktable is inconsistent with the IP address(127.10.0.1) of the network adapter, please make sure they're consistent. num(2)

Possible Cause

When verifying the device IP address, HCCL finds that the device IP address obtained on the current device is inconsistent with that configured for the current rank in the rank table. As a result, the verification fails.

For example, on rank 0, the IP address of the bound device is 127.10.0.1, but the device IP address configured for rank 0 in the rank table is 127.10.0.0. As a result, HCCL fails to verify the IP address.

Solution

Check that the device IP address configured in the rank table is consistent with the device IP address executed by each rank in the communicator.