[email protected] (Jesse 1 Robinson) writes: > Our data center folks insist on dual power feeds for everything, > sometimes infuriatingly so. To test power redundancy, they > occasionally drop one power feed or the other--with ample heads > up--and check that all devices are functioning. Other than call-home > events, we have not had any surprises so far.
when we were doing IBM's ha/cmp product ... some past posts http://www.garlic.com/~lynn/subtopic.html#hacmp we went around to various customers talking about failure modes. One customer in Manhatten had carefully chosen a building that that telco feeds from four different substations on four sides of the bldg, power from different substations on opposite sides of the bldg, and water from different water mains on opposite sides of the bldg. The datacenter was shutdown when transformer in the basement exploded and the bldg had to be evacuate because of contamination. There were a number of other customers with similar stories. It was while out talking to customers that I coined disaster survivability (to differentiate from disaster recovery) and geographic survivabilty. some past posts http://www.garlic.com/~lynn/submain.html#availability I was then asked to write a section for the corporate continuous availability strategy document. The section got pulled when both Rochester (as/400) and POK (es/9000) complained that they weren't able to meet the requirements. trivia: mainframe DB2 group were also complaining if I was allowed to proceed with (commercial) HA/CMP cluster scaleup, it would be at least 5yrs ahead of them. old post about Jan1992 meeting in Ellison's conference room http://www.garlic.com/~lynn/95.html#13 -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
