I’ve installed dozens of IBM FlashSystem storage units throughout the years and replication on about a third of them primarily for migration. But I recently had a different experience.
When attempting to create a partnership between a FlashSystem FS5200 with code level (8.4.2.1) and a FlashSystem FS5300 with code level 8.7.0.2, though we zoned them correctly via their REAL WWPNs, the GUI reported nothing was found on the fiber. Upon further review, we discovered that the system “layer” setting differed on the two units via the lssystem command as shown below:
IBM_FlashSystem:P9FS5200:superuser>lssystem |grep layer
layer storage
IBM_FlashSystem:P10FS5300:superuser>lssystem |grep layer
layer replication
This is because the default system layer changed to replication in 8.6.3. In order to establish a proper relationship the system layer setting MUST match.
More information on system layers can be found at: https://www.ibm.com/docs/en/flashsystem-c200/9.1.0?topic=system-layers
When attempting to change the original FS5200 to replication, which is the recommendation, to match the FS5300, another error was encountered about other system/cluster nodes were detected.
This is because the zoning in place, does allow them to “see” other but not effectively talk to each other. So the resolution was to stop the communication between them. This can be done by removing the zone, disabling the port(s) the new storage unit is attached to, or physically unplugging the cables designated for the replication. In our case, we removed the zone. Then run the command below to change the layer on the FS5200 to match the new FS5300 (and is the recommendation) and then zone it back.
Disabled the zone.
Executed: chsystem -layer replication
Enable the zone.
This inturn allowed us to create the partnership successfully.
