Tuesday 28th August 2018

Storage Cluster Stalled write i/o for a small number of customers

At 11:16 CEST a storage SSD started to not acknowledge write I/O anymore. Thus, any I/O with this SSD as a primary replica stalled. After the SSD was removed from operation all queued I/Os immediately completed.

Storage operation was back to normal by 11:23 CEST.