07-Oct-2003

Problems

We have been having some hangs/crashes recently on the data warehouse cluster. And while we suspect a hardware problem, there is not much information to point at a specific component.

The cluster consists of two AlphaServer GS160s soft partitioned using Galaxy software into four nodes. The individual nodes are connected with dual redundant MemoryChannel II as the cluster interconnect. We have seen virtual circuit closures, but on both the A and B memory channel adapters (?!?) and hangs on nodes in both GS160s.

Indeed a puzzle.

Posted at October 7, 2003 9:50 AM
Tag Set:

Comments are closed