Cisco: Some UCS B440 Blades Could Overheat, Need Replacing

Cisco is telling customers of its UCS blade servers that its UCS B440 model carries a risk of overheating and possible fire.

Cisco, in a Field Notice published on its Web site, wrote that failure of a MOSFET power transistor on the blade server can cause that component to overheat and emit a short flash, possibly leading to a complete failure of the circuit board. Under certain conditions, that failure could also disrupt power flow to other blades in the same chassis.

In a statement e-mailed to CRN, Cisco wrote, "As confirmed in our public Field Notice, there have been just two related incidents at customer sites since the Cisco UCS B440 blade server was introduced in May 2010. Cisco is proactively managing this situation in our normal way and with the best interests of our customers in mind. We have proactively contacted all affected customers and are discussing available options with them. This issue is specific just to B440 blade servers and no other UCS solutions are involved."

After finding two incidents of the problem in the field, Cisco determined that it needed to modify the hardware in its UCS B440 blade servers, and has launched a hardware replacement program, according to Cisco's Field Notice.

Sponsored post

The issue stemmed from a MOSFET power transistor failure which caused the component to over heat and emit a flash before it failed. Cisco said in its Field Notice called this a "random component failure," and not a system issue.

In response, Cisco wrote in its Field Notice that it is "directly contacting UCS B440 Blade Server customers and will replace UCS B440 Blade Servers currently deployed at customer sites. Cisco is making UCS B440 Blade Server hardware modifications, and a hardware replacement program has been launched. No other UCS hardware is affected."

A Cisco spokesperson declined to call this situation a "recall" of the UCS B440 server.

However, the spokesperson emailed CRN in response to an inquiry about a possible recall by writing, "I can confirm that we have proactively reached out to affected customers to discuss the contents of the Field Notice and the associated replacement program. We are talking to them about the applicability of alternative UCS solutions now, and will be providing them with replacement B440s as soon as the redesigned B440s are available."

Because a firmware update originally made to prevent the MOSFET overheating is not effective with all MOSFET failure modes, Cisco is recommending customers replace its UCS B440 M1 and UCS B440 M2 blade servers with a certain hardware version.

"A hardware replacement program has been launched and a modified UCS B440 Blade Server will be announced when it becomes available. Cisco is also working with customers on the option of alternative UCS solutions," the company wrote in its Field Notice.

Cisco's UCS (Unified Computing System) is the company's strategy for bringing servers, storage, networking, and virtualization into a single architecture managed by a unified management system.

Despite introducing its UCS blade servers less than three years ago, the company has seen strong adoption among customers. Last month, Cisco bragged about its 10,000th UCS deployment, and IDC in November reported that Cisco already has a 10.7-percent share of the fast-growing blade server market.