Lenovo Support in Indonesia After End of Life

More Memory, Slower Server

It is a genuinely counterintuitive outcome and it is common. An organisation adds memory to a ThinkSystem or System x machine, the total capacity goes up as expected, and the system runs measurably slower than it did before. Nothing has failed and no error is reported.

The Platform Has Rules About Distribution

Server memory operates across channels, and every platform specifies how modules must be distributed and matched across them. Populate correctly and all channels run in parallel at full speed. Populate unevenly — two modules where the layout calls for four or six, or mixed ranks across a channel — and the system falls back to a lower speed or an unbalanced mode to accommodate what it has been given.

It Is a Performance State, Not a Fault

This is why it goes unnoticed for years. The modules are valid, the total is right, the machine posts and boots cleanly, and no health check reports anything, because from the firmware perspective nothing is wrong. The cost is a permanent reduction in memory bandwidth that nobody attributes to the upgrade that caused it.

Buy the Set, Not the Shortfall

The fix is to specify against the channel layout rather than against the capacity gap — adding the set the platform requires instead of the number of modules that reaches the total. ICD works from the serial to the installed population and quotes the correct matched set, across ThinkSystem and System x generations past vendor support, tested and covered for one to three years.