Cisco warns of hardware failure in DIMM memory components

A manufacturing defect affecting some DIMMs made in late 2020 could cause persistent memory errors and server failure. Cisco recommends customers replace impacted DIMMs.

Cisco has warned that some of its dual in-line memory modules are failing prematurely due to a manufacturing error.

Customers using specific Cisco DIMMs that appear in products such as the Cloud Native Broadband Router (CNBR) and the Unified Computing System (UCS) server are encouraged to replace the component to prevent server failures. The DIMMs with faulty chips are 16 GB, 32 GB and 64 GB components manufactured during the second half of 2020.

Customers might notice a faulty DIMM from persistent correctable memory errors. However, Cisco discouraged customers from relying solely on that metric, as operating system and reliability, availability and service (RAS) features might mask some errors.

Instead, the company recommended customers use its Serial Number Validation Tool to test whether their hardware uses affected DIMMs. Not all DIMMs contain the manufacturing error.

Companies that fail to replace the DIMM might discover that error messages disappear in systems after a repair even if issues remain. Over time, these memory errors can lead to uncorrectable memory errors and an unexpected server reset. An uncorrectable memory error can reduce the total available memory.

Products with faulty DIMMs include models of Cisco's Cloud Services Platform, the CNBR and the UCS. The flaw also affects Business Edition 6000 and 7000 VoIP systems. A complete list of impacted product IDs is available in Cisco's field notice on Friday.

Companies can order replacement parts through Cisco. Customers in South America, Central America, Asia and non-EU countries should expect shipping delays of three months. The replacement DIMM may not immediately show as healthy, particularly if the physical DIMM seating is incorrect. Cisco recommends customers run server memory diagnostics to minimize early runtime errors.

In other Cisco news last week, the company introduced a predictive analytics engine for network management. It uses AI to anticipate network health issues before they occur based on previous network behavior.

Cisco developed the engine using its data over the last two years. The SaaS product will ingest customer data to tailor predictions to users' specific networks.

Cisco plans to unveil the engine at Cisco Live in June. The company plans to integrate the service into other Cisco products in the coming years.

The engine has shown promise at accurately predicting when network traffic paths will degrade in early customer trials. However, IT professionals will still have to learn about manufacturing defects the traditional way.

Madelaine Millar is a news writer covering network technology at TechTarget. She has previously written about science and technology for MIT's Lincoln Laboratory and the Khoury College of Computer Sciences, as well as covering community news for Boston Globe Media.

Dig Deeper on Cloud and data center networking

Unified Communications
Mobile Computing
Data Center