Help Center/ ModelArts/ Resource Management/ Elastic Cluster/ Abnormal Status of a Dedicated Resource Pool

Updated on 2024-04-30 GMT+08:00

View PDF

Abnormal Status of a Dedicated Resource Pool

Resource Quota Limit

When you use a dedicated resource pool (for example, scaling resources, creating a VPC, creating a VPC and subnet, or interconnecting a VPC), if the system displays a message indicating that the resource quota is limited, submit a service ticket.

Creation Failed/Change Failed

Log in to the ModelArts management console. In the navigation pane, choose Dedicated Resource Pools > Elastic Cluster.
Click Records on the right of Create. On the Records dialog box, view failed task records.
Figure 1 Creating a resource pool failed
Hover the cursor over , view the cause of task failures.

By default, failed task records are sorted by application time. A maximum of 500 failed task records can be displayed and retained for three days.

Locating Faulty Node

ModelArts will add a taint on a detected K8S faulty node so that jobs will not be affected or scheduled to the tainted node. The following table lists the faults can be detected. You can locate the fault by referring to the isolation code and detection method.

**Table 1** Isolation code
Isolation Code	Category	Sub-Category	Description	Detection Method
A050101	GPU	GPU memory	GPU ECC error exists.	Run the nvidia-smi -a command and check whether Pending Page Blacklist is Yes or the value of multi-bit Register File is greater than 0. For Ampere GPUs, check whether the following content exists: Uncorrectable SRAM error Remapping Failure records Xid 95 events in dmsg (For details, see NVIDIA GPU Memory Error Management.) The Ampere architecture has the following levels of GPU memory errors: L1: These are single-bit ECC errors that can be corrected. They do not affect the running services. To check for these errors, run the nvidia-smi -a command and look for Volatile Correctable. L2: These are multi-bit ECC errors that cannot be corrected. They cause the running services to fail and require a process restart to recover. To check for these errors, run the nvidia-smi -a command and look for Volatile Uncorrectable. L3: These are unsuppressed errors and may affect other services. They require a card reset or a node reboot to clear. To check for these errors, look for the Xid events that contain the number 95. (The Remapped Pending records are only for reference. You need to reset the cards when the service is idle to trigger the remapping process.) L4: These are errors that require a card replacement. To check for these errors, look for the SRAM Uncorrectable field that is greater than 4 or the Remapped Failed field that is not zero.
A050102	GPU	Other	The nvidia-smi output contains ERR.	Run nvidia-smi -a and check whether the output contains ERR. Normally, the hardware, such as the power supply or the fan, is faulty.
A050103	GPU	Other	The execution of nvidia-smi times out or does not exist.	Check that exit code of nvidia-smi is not 0.
A050104	GPU	GPU Memory	ECC error occurred 64 times.	Run the nvidia-smi -a command, locate Retired Pages, and check whether the sum of Single Bit and Double Bit is greater than 64.
A050148	GPU	Other	An infoROM alarm occurs.	Run the nvidia-smi command and check whether the output contains the alarm "infoROM is corrupted".
A050109	GPU	Other	Other GPU errors	Check whether other GPU error exists. Normally, there is a faulty hardware. Contact the technical engineer.
A050147	IB	Link	The IB NIC is abnormal.	Run the ibstat command and check whether the NIC is not in active state.
A050121	NPU	Other	A driver exception is detected by NPU DCMI.	The NPU driver environment is abnormal.
A050122	NPU	Other	The NPU DCMI device is abnormal.	The NPU device is abnormal. The Ascend DCMI interface returns a major or urgent alarm.
A050123	NPU	Link	The NPU DCMI net is abnormal.	The NPU network connection is abnormal.
A050129	NPU	Other	Other NPU errors	Check whether other NPU error exists. You cannot rectify the fault. Contact the technical engineer.
A050149	NPU	Link	Check whether the network port of the hccn tool is intermittently disconnected.	The NPU network is unstable and intermittently disconnected. Run the hccn_tool-i ${device_id} -link_stat -g command and the network is disconnected more than five times within 24 hours.
A050951	NPU	GPU memory	The number of NPU ECCs reaches the maintenance threshold.	The NPU's HBM Double Bit Isolated Pages Count value is greater than or equal to 64.
A050146	Runtime	Other	The NTP is abnormal.	The ntpd or chronyd service is abnormal.
A050202	Runtime	Other	The node is not ready.	The node is unavailable. The K8S node contains one of the following taints: node.kubernetes.io/unreachable node.kubernetes.io/not-ready
A050203	Runtime	Disconnection	The number of normal AI cards does not match the actual capacity.	The GPU or NPU is disconnected.
A050206	Runtime	Other	The Kubelet hard disk is read-only.	The /mnt/paas/kubernetes/kubelet directory is read-only.
A050801	Node management	Node O&M	Resource is reserved.	The node is marked as the standby node and contains a taint.
A050802	Node management	Node O&M	An unknown error occurs.	The node is marked with an unknown taint.
A200001	Node management	Driver upgrade	The GPU is being upgraded.	The GPU is being upgraded.
A200002	Node management	Driver upgrade	The NPU is being upgraded.	The NPU is being upgraded.
A200008	Node management	Node admission	The admission is being examined.	The admission is being examined, including basic node configuration check and simple service verification.
A050933	Node management	Fault tolerance Failover	The Failover service on the tainted node will be migrated.	The Failover service on the tainted node will be migrated.
A050931	Training toolkit	Pre-check container	A GPU error is detected in the pre-check container.	A GPU error is detected in the pre-check container.
A050932	Training toolkit	Pre-check container	An IB error is detected in the pre-check container.	An IB error is detected in the pre-check container.