Community
Participate
Working Groups
We encounter ZK write errors through concurrent write access on the ZK node holding the repository assignment information. The reason for the too frequent access to this nodes is that too much information is stored in one ZK node. So it need to be splitted in seperate nodes To solve this problem, we need to do: 1. summarize the current repo assignment data stored in ZK 2. propose a ZK tree/node structure to store splitt this data into seperate nodes 3. consider the option to remove unnecessary information from ZK, if possible
Konrad, please feel free to post any Gerrit reviews you'd like to have feedback on.