IBM Support in Jordan After End of Life

IBM End-of-Life Support in Jordan

A growing share of the IBM hardware running in Amman was not bought to serve Jordanian users. It was bought to stand behind a primary platform somewhere else in the region — a recovery site, a replication target, a second copy that exists so a regulator or a board can be told that one exists. That purpose changes what end of support means, and it changes it in a direction most support planning gets backwards.

Standby hardware hides its faults

A production server tells you when something is wrong. Load makes faults surface: a degrading drive slows a query, a failing supply trips an alert, a marginal DIMM throws a correctable error someone notices. A standby platform does none of this. It sits at near-zero utilisation, its components age on the calendar rather than under load, and the first honest test of its condition is the moment it is asked to take over. That is a poor moment to discover that an enclosure power supply died eight months ago in a redundant pair.

The failover test is only as good as the part behind it

Organisations that exercise their recovery site properly find these faults, which is the point. What determines whether finding them is useful is whether the component can be replaced. On an IBM generation that has passed withdrawal, the honest answer is often no — not quickly, and not through the vendor. So the test produces a finding, the finding produces a risk register entry, and the risk register entry sits there because there is no supply route to close it. A maintenance agreement over withdrawn Power and Storwize hardware is what converts that entry back into an action.

Generational mismatch between the two sites

Recovery platforms are rarely the twin of what they protect. They are bought later, or bought to a tighter budget, or inherited when the primary was refreshed. The practical result is that the two sites run different Power generations with different feature codes, and a spare that fits the primary will not necessarily fit the standby even where the model numbers rhyme. ICD matches parts to the feature code and the actual configuration rather than the model name, because on a recovery platform a part that does not fit is indistinguishable from no part at all.

What this means for the contract

The support position for a recovery site should be written against the failure modes of idle hardware, not busy hardware: power supplies in redundant pairs where a single failure is invisible, SAS cabling between enclosures, and drives in replication targets that have been spinning without being read. Those are the features worth holding locally in Amman. Transit is three to four days from Cairo on the Levant lane, with clearance typically adding one to three days — which for a platform that exists to be available immediately is the whole argument for holding stock rather than ordering it. Browse the IBM server parts catalogue, or the wider Jordan server parts page.