Yuan Wang

``` createSDSConfig("masterauth", xxx, EMPTY_STRING_IS_NULL, server.masterauth, xxx), createSDSConfig("requirepass", xxx, EMPTY_STRING_IS_NULL, server.requirepass, xxx), ``` I checked the code, `masterauth` and `requirepass` configurations have `EMPTY_STRING_IS_NULL`, so it is ok if they set empty,...

Thanks @ddmg-250 we are enhancing data migration now

@ymcyyds you could have a look at [Add Atomic Slot Migration (ASM) support #14414](https://github.com/redis/redis/pull/14414), it may help you

maybe #13495 fixed this issue, before that commit, the replica doesn't process any request when flushing old data, please have a look. btw, you maybe also perform `config set replica-lazy-flush...

@wstar05 may i know under what condition you find this issue? why is `getpeername` system call failed and then successful @smartlee i remember you encountered the similar problem?

Hi @wstar05 i have a question, if failing to call `getpeername`, we won't update node ip info, as code https://github.com/redis/redis/blob/unstable/src/cluster_legacy.c#L2264

i was thinking if 30s is enough, as we know, tcp keepalive works on the kernel thread instead of application thread, if tcp keepalive causes this failure, it means the...