Binbin
yes, the zlibc_trim now become a dead code (i will remove it after the code become solid), considering the problem happening in https://github.com/redis/redis/pull/13133#discussion_r1532415573, i kind of want to remove lua_close...
> When a primary has lost a slot, it deletes all the keys and replicates DEL to its replicas? That's why there is some commands to replica even when write...
@zuiderkwast thanks, good catch with the manual failover one, i added a new test that can confirm it.
Thinking of backporting it to 7.2? I added the label, let me know if we don't want to backport it.
We don't seem to expose in info that the server may be in pause state and its purpose, do we need it?
> Has there been a report that this issue occurred on 7.2? Nope
> Should we remove it from backport? It's a bit painful to cherry pick into 7.2 as it doesn't have the cluster split code. ohh, i guess, if there is...
later, i might add some simple tcl tests later to cover this path to make sure the order is consistent
Ci link: https://github.com/valkey-io/valkey/actions/runs/10913083514/job/30288951861#step:8:5121
@madolson @zuiderkwast do you guys want to take a look with this?