Hi everyone,
I have a Redis server (version 3.2.6) that recently started crashing whenever the RDB is saved. I've temporarily disabled the automatic save to keep the server running, but all bgsave and bgrewriteaof commands fail. Unfortunately AOF is not enabled.
The memory used is 10 GB on a 48 GB Linode instance. The last successful RDB is about 3 GB on disk. When saving now, it writes a temp file up to about 2 GB before crashing. Running bgrewriteaof also creates an incomplete file before crashing.
I know 3.2.6 is old. I would be fine running a blocking save instead of bgsave but I'm concerned that will only crash the main process, losing all of the changes since the last RDB.
My guess is that there is either corruption with some of the newer keys, or bad memory in the forked process. There is no slave but I also assume that setting one up would crash while doing a full sync.
Any tips for recovering from this problem without losing any data currently in memory? Thanks!
— Manton