When the database restarts, it tries to load the entire WAL into memory for compression, which is evidently not the correct method. Even though the memory has been upgraded to 200GB, there is still insufficient memory to initiate the startup process. Attempting to rebuild the index, downgrade to versions 2.0, 2.2, or 2.4, restore from an image, and so on, have not addressed the problem. It is unfortunate that despite the desire to manage memory usage, there are no corresponding configuration settings in the configuration file. Is there an alternative approach that can resolve this issue?
1 Like
